Customer Service Outsourcing · independent research
INDEPENDENT RESEARCH · UPDATED OCTOBER 2026

The Best Customer Service Outsourcing Companies in 2026

Customer service outsourcing is the practice of running a company's customer-facing operations through an external provider: inbound and outbound voice, live chat, email and messaging, technical support, back office processing, and the quality and workforce management that holds them together, delivered onshore, nearshore or offshore. This report evaluates twelve providers against eight weighted criteria covering service quality, talent model, delivery footprint, security posture, scalability, cost-to-quality, automation capability and speed to launch, drawing on SEC and exchange filings, analyst assessments, government statistics and independent survey research.

No paid placements. No sponsored rankings. No provider has paid, or can pay, to be included, ranked higher, described differently, or removed. This publication accepts no advertising, no affiliate compensation and no referral fees from any ranked provider. Every external figure links to its primary source at the point it is used.

12
providers scored
8
weighted criteria
19
buying scenarios
$0
paid by any provider
No paid placementsNo sponsored positionsNo affiliate or referral feesEvery external claim links to its primary source
Editorial independence

What this page is, and what it is not

Three kinds of page rank customer service outsourcing. The vendor blog that ranks itself first. The directory that sells position and calls the price a listing fee. The affiliate page earning a commission on the click. All three are legitimate businesses and none of them are research. This page is the fourth kind, and the only way to show that is to publish the method and the facts that work against its own conclusion — starting with the fact that the provider ranked first is the smallest company on the page and appears in no analyst assessment at all.

Do you accept paid placements or sponsorships?

No. No provider on this page has paid to appear, to rank higher, to be described differently or to be removed, and there is no mechanism by which they could. This publication carries no advertising, no affiliate links and no sponsored slots, and takes no referral fee from any ranked provider.

Is the ranking pay-to-play?

No. Providers were not invited to participate, were not shown their scores in advance, and were not given an opportunity to respond before publication. Several would object to what is written about them here, which is the point. A ranking a vendor could veto is a directory listing, not research.

Why is each provider included?

Category fit and buyer relevance. The set deliberately mixes tiers, because a ranking of only large incumbents is useless to anyone not buying at enterprise scale, and a ranking of only boutiques has no reference point. Inclusion is not an endorsement and exclusion is not a judgment — it usually means the provider sits outside the category this page covers.

What are the scores, exactly?

Editorial judgments applied consistently against the published rubric, using public evidence. They are not audited operational figures, and nobody has inspected a delivery floor. The full per-criterion matrix is published so you can disagree with any individual judgment and recompute the order with your own weights.

There is no separate corrections panel on this page. The facts that work against the top-ranked provider sit where a buyer meets them: its absence from Everest Group's PEAK Matrix in both 2025 and 2026 is on its provider card, in the facts sheet, in the evidence table and in its own FAQ question; the four criteria it loses outright are named in a panel on the scoring matrix; the contradiction between its marketing and its live Clutch rating is on the card; and the four situations where it is the wrong choice each route to a named alternative.

Methodology

How the twelve companies were scored

Most customer service outsourcing rankings are size rankings with a methodology section attached, which is why the same four names lead every list and why the list is useless to anyone not buying tens of thousands of seats. This rubric bets differently: that what decides whether a programme hits its numbers in year one and survives to year three is quality discipline, agent tenure and value per dollar, not seat count. Scalability therefore carries 12% rather than 40%. If you weight scalability and footprint at 40% instead, Teleperformance wins this ranking and you should buy Teleperformance. The full matrix is printed below so you can do that arithmetic rather than take this page's word for it.

What the composite score is made of

Each criterion's share of the weighted total. Weights are published here, ahead of the ranking.

20%15%15%14%12%12%7%5%Share of the composite score
Service quality & QA disciplineTalent model & attrition controlDelivery footprint & language coverageSecurity & compliance postureScalability & enterprise readinessCost-to-quality balanceAI & automation capabilitySpeed to launch & contract flexibility
Weights are editorial and published here, ahead of any result, so the ranking reads as a method rather than an opinion. Service quality carries the most because it is the outcome the programme exists to move; scalability carries 12% because it is decisive above about a thousand seats and largely irrelevant below five hundred.
Show the data
Criterion~WeightWhat it measures
Service quality & QA discipline20%Published outcomes, QA sampling depth, calibration cadence, and whether quality is tied to commercial terms rather than to a slide.
Talent model & attrition control15%Dedicated versus pooled staffing, tenure posture, hiring geography, and whether product knowledge stays in the account.
Delivery footprint & language coverage15%Countries, sites, timezones and verified languages; onshore, nearshore and offshore mix and continuity redundancy.
Security & compliance posture14%ISO 27001 and 27701, SOC 2, PCI DSS with its level, ISO 42001 — separating certified from compliant from claimed, and checking scope.
Scalability & enterprise readiness12%Seat capacity, procurement maturity, analyst-tier recognition and Fortune 500 density.
Cost-to-quality balance12%Value per dollar rather than lowest rate, judged against disclosed delivery geography and overhead structure.
AI & automation capability7%Productised automation in production, measured containment, and data capability — as distinct from roadmap.
Speed to launch & contract flexibility5%Ramp time, minimum commitment, pilot availability, lock-in and exit terms.

The eight criteria and what each one measures

Weighted criteria, definitions and share of the composite score. Weights were set before any provider was scored.
CriterionWeightWhat it measures and how it was judged
Service quality & QA discipline20%Published outcomes, QA sampling depth, calibration cadence, and whether quality is tied to commercial terms rather than to a slide.
Talent model & attrition control15%Dedicated versus pooled staffing, tenure posture, hiring geography, and whether product knowledge stays in the account.
Delivery footprint & language coverage15%Countries, sites, timezones and verified languages; onshore, nearshore and offshore mix and continuity redundancy.
Security & compliance posture14%ISO 27001 and 27701, SOC 2, PCI DSS with its level, ISO 42001 — separating certified from compliant from claimed, and checking scope.
Scalability & enterprise readiness12%Seat capacity, procurement maturity, analyst-tier recognition and Fortune 500 density.
Cost-to-quality balance12%Value per dollar rather than lowest rate, judged against disclosed delivery geography and overhead structure.
AI & automation capability7%Productised automation in production, measured containment, and data capability — as distinct from roadmap.
Speed to launch & contract flexibility5%Ramp time, minimum commitment, pilot availability, lock-in and exit terms.

What these scores are, and what they are not

These are editorial judgments applied consistently against the published rubric using public evidence. They are not audited operational figures and nobody from this publication has inspected a delivery floor. Six of these twelve providers disclose no revenue at all and two publish no headquarters city, and those absences are scored as the uncertainty they represent rather than filled with estimates. The margin between first and second is 0.66, which sits inside this method's own 0.4 to 0.7 target band — the only page in this network where it does.

Composite weighted score, all twelve providers

Zero-baselined 0–10 scale. Full per-criterion detail sits in the scoring matrix.

Helpware9.03Concentrix8.37TELUS Digital8.34Sutherland8.12ibex8.11IntouchCX8.07Teleperformance8.05Foundever7.94Transcom7.90Alorica7.83Startek / CCI Global7.71HGS7.68
Computed from the published weights and the per-criterion scores in the scoring matrix. Zero-baselined rather than truncated. The spread from first to twelfth is 1.35 points — the narrowest in this network — and the eleven providers below first place are separated by less than seven tenths of a point from each other. Scenario fit matters considerably more than ordinal position here.
Show the data
Provider~Composite
Helpware9.03
Concentrix8.37
TELUS Digital8.34
Sutherland8.12
ibex8.11
IntouchCX8.07
Teleperformance8.05
Foundever7.94
Transcom7.90
Alorica7.83
Startek / CCI Global7.71
HGS7.68
Market context

Five numbers that should change how you buy in 2026

Each argument below follows the same shape: what the data says, where it contradicts itself or the prevailing story, and what that means for a decision you are about to make. Every figure links to its primary source at the point of use, and where no independent source exists that is stated rather than filled.

The incumbents are writing off billions while the forecasts say the market compounds at 13%

Grand View Research values customer experience BPO at $113.0 billion in 2025 rising to $296.3 billion by 2033, a compound rate of 13.0%. Against that, Concentrix FY2025 results recorded a $1,523.3 million goodwill impairment in Q4 FY2025 and its Q3 FY2026 results a further $1,050.0 million in Q3 FY2026 — $2.57 billion in four quarters, the second tranche explicitly justified by the company's own share price and market capitalisation.

Teleperformance's FY2025 results posted FY2025 revenue down 0.7% and its H1 2026 results down 4.5%, four consecutive declining periods. a $500 million goodwill impairment was booked against the TELUS digital experience unit months before TELUS completed the privatisation of TELUS Digital at US$4.50 a share. ISG Index recorded total BPO annual contract value down 14% in FY2025 to its lowest since 2020.

What this means for you. Scale is not protecting the incumbents, and the market disbelieves the carrying value of the two largest. That is leverage: this is a good year to negotiate rate, term and exit provisions, and a bad year to sign a long lock-in with a provider whose balance sheet you have not read.

Goodwill written off by the listed incumbents

US$ millions. Impairments disclosed in the last four reported quarters.

042785312801706$1,523MConcentrixQ4 FY2025$1,050MConcentrixQ3 FY2026$500MTELUS, digitalexperience unit~$105MTeleperformanceFY2025US$ millions
Sources: Concentrix FY2025 results for the $1,523.3 million Q4 FY2025 impairment and its Q3 FY2026 results for a further $1,050.0 million in Q3 FY2026, the second explicitly attributed to the company's own share price and market capitalisation; a $500 million goodwill impairment booked against the TELUS digital experience unit on a 30 June 2025 test and flagged as a critical audit matter; and Teleperformance's FY2025 results, which recorded €97 million against goodwill and intangibles, converted here at approximately $105 million and marked as converted in the data table. Note: Foundever, Alorica, Sutherland, IntouchCX, Transcom and Startek are private and disclose no impairment data, so their absence from this chart is a disclosure gap rather than evidence of health.
Show the data
Impairment~US$ millionsPeriod and basis
Concentrix1,523.3Q4 FY2025, goodwill
Concentrix1,050.0Q3 FY2026, goodwill
TELUS digital experience unit500.0Tested 30 Jun 2025
Teleperformance~105.0FY2025, €97M converted

Analyst standing does not predict financial health

Everest Group's CXM Services PEAK Matrix named Foundever a Leader in both 2025 and 2026. Eight months later Foundever's August 2026 recapitalisation cut nearly $900 million of debt, took a $225 million rescue equity injection from its majority shareholders, and lost both its group chief executive and its executive chairman with effect from 31 July 2026.

Alorica was promoted to Leader in the 2026 Americas matrix while disclosing no financials at all. TELUS Digital retains Leader status for an eighth year as a dark subsidiary publishing no revenue, headcount or country data since 2024. Startek vanished from the 2026 matrix entirely, assessed only on a pre-merger basis after it combined with CCI Global in April 2026.

What this means for you. The PEAK Matrix measures capability and market impact as assessed through a provider's own briefing process. It is a useful signal about capability and a poor one about solvency. If your procurement gates on analyst tier, understand what the gate is actually testing, and read the financials separately.

Scale varies by two orders of magnitude, and so does what your account is worth

Teleperformance disclosed 446,716 people at the end of 2025. HGS FY2026 results disclosed 17,110. Helpware discloses roughly 4,000. These are all described as customer service outsourcing providers and they are not comparable businesses.

The arithmetic that follows is simple and rarely stated. A 200-seat programme is about 0.04% of a 446,000-person provider's headcount and roughly 5% of a 4,000-person one. The first gets a shared account manager and a standard playbook; the second gets named leadership because it has to. Neither is wrong — it depends entirely on whether you need a provider's scale or its attention.

What this means for you. Work out which of those two you are buying before you shortlist, because the answer eliminates most of this list immediately. If you need 5,000 seats across twenty languages, the boutiques cannot serve you at any price. If you need 200 seats and real attention, the incumbents structurally cannot give it to you.

Headcount across the ranked providers

Thousands of people, most recent disclosed figure. Scale is what this category sells, and it varies by two orders of magnitude.

Teleperformance446,716Foundever~130,000Alorica100,000+TELUS Digital78,879Startek / CCI Global50,000+Sutherland40,000+ibex~35,000Transcom30,000+IntouchCX25,000+HGS17,110Helpware4,000+
Most recent disclosed figures: Teleperformance at 31 December 2025; TELUS completed the privatisation of TELUS Digital at 31 December 2024, its last standalone disclosure before delisting; HGS FY2026 results at 31 March 2026; ibex FY2026 results for FY2026; Foundever's August 2026 recapitalisation and Alorica from company statements; Startek and CCI Global completed their merger post-merger; IntouchCX from a February 2024 statement that is now more than two years old. Concentrix discloses no current headcount in its releases and is therefore absent from this chart rather than estimated. Helpware is the smallest provider ranked here by a factor of four, which is a fact its first place on this rubric does not change.
Show the data
Provider~HeadcountAs of
Teleperformance446,71631 Dec 2025
Foundever~130,000Company statement, 2026
Alorica100,000+Company statement, 2026
TELUS Digital78,87931 Dec 2024, last standalone
Startek / CCI Global50,000+Post-merger, Apr 2026
Sutherland40,000+Company statement, undated
ibex~35,000FY2026
Transcom30,000+Owner statement
IntouchCX25,000+Feb 2024 — more than two years old
HGS17,11031 Mar 2026
Helpware4,000+Company-disclosed
ConcentrixNot disclosedNo current figure published

Wage arbitrage is real, and it is not what you are billed

Site Selection Group puts agent base hourly wages at $17–22 in the United States, $5–9 in Eastern Europe, $4–7 in Latin America and the Caribbean, $2–3 in the Philippines and $1.50–2.50 in India.

Those are wages. The gap between a wage and your invoice is the provider's facilities, management, technology, quality function and margin, and it varies enormously across the providers on this page — a 446,000-person incumbent carries overheads a 4,000-person provider does not, and a 4,000-person provider lacks procurement leverage the incumbent has. No independent source publishes billable rates by region, and every figure in circulation traces to provider marketing.

What this means for you. Do not negotiate against a rate benchmark you read somewhere, because it does not exist. Negotiate against your own cost per resolved contact, and ask any provider quoting a nearshore rate to show you the delivery location it assumes.

Agent base hourly wages by delivery region

US dollars per hour. These are wages paid to agents, not the rates you will be billed.

United States$17–22Eastern Europe$5–9Latin America / Caribbean$4–7Africa$1.50–3.75Philippines$2–3India$1.50–2.50$0$6$12$18$24
Source: Site Selection Group, 2024. These are wages, not billable rates. The distinction matters because the difference between them is the provider's entire cost structure — facilities, management, technology, quality and margin — and it varies enormously between a boutique and a 446,000-person incumbent. Note: no independent source publishes billable hourly rates by region. Every "$12–18 nearshore" figure in circulation traces to provider marketing, so this page publishes the independently sourced wages and states the distinction rather than republishing a vendor number.
Show the data
Region~Wage bandSource
United States$17–22Site Selection Group, 2024
Eastern Europe$5–9Site Selection Group, 2024
Latin America / Caribbean$4–7Site Selection Group, 2024
Africa$1.50–3.75Site Selection Group, 2024
Philippines$2–3Site Selection Group, 2024
India$1.50–2.50Site Selection Group, 2024

AI is restructuring the supplier side faster than it is replacing agents

Gartner predicts agentic AI will autonomously resolve 80% of common customer service issues by 2029 and Salesforce State of Service expects half of all cases AI-resolved by 2027. ContactBabel's 2026 US Decision-Makers' Guide measures 18% of web chats handled with no human agent, up from 6% in 2020, with roughly 22% still escalating.

The measurable effect so far is on suppliers rather than on agent headcount. Foundever's lenders cited clients insourcing with AI. ISG Index attributes falling contract value to AI compressing what each deal is worth while award counts hold. US Bureau of Labor Statistics projects US customer service representative employment down 5% by 2035, already down from 2,858,710 in May 2023 to 2,666,000 in 2025. And Gartner found 64% of consumers would prefer companies did not use AI for service at all.

What this means for you. Ask every provider what proportion of its revenue sits with clients who have announced AI-first service strategies, and write an exit clause you have actually tested. Vendor stability is a selection criterion in this category now, not a footnote.

AI containment: projected against measured

Percent of contacts. The first two are forecasts; the third and fourth are measurements.

02245679080%Gartner: agentic AIresolves by 202950%Salesforce: casesAI-resolved by 202718%Measured: chats withno human agent22%Measured: chats stillescalatingPercent
Sources: Gartner for 80% of common issues autonomously resolved by 2029; Salesforce State of Service, surveying 6,500 service professionals across 40 countries, for 50% of cases AI-resolved by 2027; and ContactBabel's 2026 US Decision-Makers' Guide measuring 18% of web chats handled with no human agent, with ContactBabel's 2025 edition finding roughly 22% still escalating to another channel. Gartner separately found 64% of 5,728 consumers would prefer companies did not use AI for customer service and 53% would consider switching over it. The commercial consequence is already visible: Foundever's August 2026 recapitalisation cut nearly $900 million of debt with lenders citing clients building support in-house with AI rather than outsourcing it.
Show the data
Measure~PercentType and source
Agentic AI resolving common issues by 202980Projection, Gartner
Cases AI-resolved by 202750Projection, Salesforce (n=6,500)
Web chats handled with no human agent, 202618Measured, ContactBabel (n=207)
Chats still escalating to another channel22Measured, ContactBabel
$2.57B
goodwill impaired by one provider in four quarters
Concentrix filings
446,716
people at the largest provider
Teleperformance, 31 Dec 2025
−14%
BPO contract value, FY2025
ISG Index
6 of 12
providers disclose no revenue at all
This page's facts sheet

Baseline your own numbers before you shortlist anyone

Most of the decisions on this page get made before a vendor is contacted. Take the ten metrics in the KPI baseline set, fill in what your operation does today, and only then read the ranking — a provider is a good fit relative to a number you already know, not in the abstract. Nothing here is gated and there is no email form on this page.

The ranking

The 12 best customer service outsourcing companies in 2026

Every card uses the identical template: rank, name, role label, composite score, status badges, a short narrative, strengths, limitations, and who each provider is and is not for. Four of these companies changed ownership or listing status in the last two years, two wrote off more than a billion dollars of goodwill, and one lost both its chief executive and its chairman. None of that is softened, and nor is the top-ranked provider's own record.

1

Helpware

Editor's pick · Best for mid-market attention
9.03Composite
SOC 2 Type II, company-confirmedISO 27001 & ISO 9001HIPAA & GDPR compliantAbsent from Everest PEAK MatrixSmallest provider ranked hereClutch 4.8 / 47 · G2 4.9 / 30

Founded 2015 in Lexington, Kentucky. Roughly 4,000 people across 19 locations in 11 countries in 45+ languages — the smallest provider here by a factor of four. It ranks first because the rubric weights service quality, tenure and cost-to-quality rather than seat count, and it loses footprint and scalability to Teleperformance by the widest margins in this matrix.

Strengths
  • Scores 9.6 on cost-to-quality and 9.4 on service quality and speed to launch — the highest marks on each.
  • The only provider here with a verified review rating: Clutch 4.8 / 47 and G2 4.9 / 30.
  • 45+ languages across four continents, blending onshore, nearshore and offshore on one contract.
  • Named operational leadership on 20–500 seat programmes, which the incumbents cannot offer.
  • Vendor-reported: 90% CSAT, 86% employee satisfaction.
Limitations
  • Absent from Everest Group's CXM PEAK Matrix in 2025 and 2026 at every tier, where six providers here are Leaders. If your procurement gates on analyst standing, this is disqualifying.
  • At roughly 4,000 people it is a hundredth the size of Teleperformance, with no evidence of a 5,000-seat multi-country consolidation.
  • Revenue not disclosed; headcount company-disclosed rather than audited.
  • Its own marketing overstates its Clutch rating, citing 5.0 across 150 reviews against a live profile of 4.8 across 47.
  • Its public compliance page shows SOC 2 Type 1 on a software-division URL, contradicting the Type II it confirms.
  • Meaningful delivery capacity sits in Ukraine — a continuity consideration.
Best fitProgrammes of 20–500 seats wanting named leadership and multilingual blended delivery without being a rounding error.
Not best forConsolidations above about 1,000 seats, analyst-gated procurement, or buyers needing audited vendor financials.
Why it is includedThe only provider here for whom a mid-market programme is a significant account.
2

Concentrix

Best for global enterprise consolidation
8.37Composite
$2.57B impaired in four quartersFY2026 operating loss guidedEverest Leader 2025 and 2026NASDAQ: CNXC70+ markets

Headquartered in Newark, California, listed on NASDAQ, operating across 70-plus markets with 2,000-plus clients including 160-plus of the Fortune Global 500. FY2025 revenue of $9,825.8 million, up 2.2%. It is the largest provider here by revenue and the one the market has most visibly repriced.

Strengths
  • Scores 9.8 on both delivery footprint and scalability, the joint-highest in this matrix alongside Teleperformance.
  • Everest Group CXM Leader in both 2025 and 2026, and the only dual Leader and Star Performer holder in 2025.
  • FY2025 revenue of $9,825.8 million, up 2.2%, audited and filed.
  • 2,000-plus clients including 160-plus Fortune Global 500 — unmatched procurement maturity.
  • Scores 9.4 on security, the highest here.
Limitations
  • $2.57 billion of goodwill impaired in four reported quarters — $1,523.3 million in Q4 FY2025 and $1,050.0 million in Q3 FY2026, the second explicitly attributed to its own share price and market capitalisation.
  • FY2025 net loss of $1,278.9 million, and nine-month FY2026 net loss of $911.2 million. FY2026 guidance now implies an operating loss of $512–522 million.
  • Scores 5.2 on speed to launch and 6.4 on cost-to-quality, the weakest on both after Teleperformance.
  • Its own 10-K states the industry is characterised by high staff attrition.
  • Discloses no current headcount figure in its releases.
Best fitGlobal enterprise consolidations across many markets and languages where procurement maturity and footprint are the binding constraints.
Not best forMid-market programmes, short pilots, or buyers who read the balance sheet before the capability deck.
Why it is includedIt is the scale benchmark for this category and the clearest illustration that scale is not protecting incumbents.
3

TELUS Digital

Everest Leader, now a dark subsidiary
8.34Composite
Delisted 31 Oct 2025$500M impairment before buyoutEverest Leader, 8th yearNo data published since 2024

Headquartered in Vancouver. TELUS completed the privatisation of TELUS Digital on 31 October 2025 at roughly US$539 million, taking it to 100% ownership and delisting it from the NYSE and TSX. Its last standalone disclosure was FY2024: revenue of $2,658 million, down 2%, with a net loss of $61 million and 78,879 staff.

Strengths
  • Everest Group CXM Americas Leader for an eighth consecutive year, and a Major Contender in EMEA and APAC.
  • Scores 9.2 on AI capability, the highest in this matrix, reflecting genuine digital and data engineering depth.
  • 78,879 staff at its last disclosure with established global delivery.
  • Full TELUS ownership removes the minority-shareholder tension that preceded the buyout.
Limitations
  • It publishes nothing. No revenue, headcount or country data since FY2024. Any comparison citing its operating scale is relying on figures at least 21 months old.
  • TELUS booked a $500 million goodwill impairment against the digital experience unit on a 30 June 2025 test — months before completing the buyout at US$4.50 a share — and flagged it as a critical audit matter.
  • Its last reported year showed revenue down 2% and a net loss of $61 million.
  • Scores 5.4 on speed to launch, among the lowest here.
  • No verified review-platform rating with a published count.
Best fitEnterprise buyers wanting digital engineering and AI capability alongside service delivery, from a provider inside a large telecommunications group.
Not best forAnyone whose diligence requires current financial disclosure, or who needs to verify operating scale independently.
Why it is includedIt is the clearest case on this page of analyst standing and financial disclosure pointing in opposite directions.
4

Sutherland

Best engineering-led transformation
8.12Composite
Private, no financialsEverest Major Contender40,000+ staff, 70+ officesNo HQ city published

Founded 1987, with 40,000-plus staff across 70-plus offices serving clients in 144-plus countries. It is engineering-led rather than seat-led, and is the most consistently placed Major Contender in this set without being a Leader.

Strengths
  • Scores 9.0 on both AI capability and scalability, among the strongest here outside the two largest incumbents.
  • Engineering and automation depth genuinely integrated into service delivery rather than sold alongside it.
  • 70-plus offices with clients in 144-plus countries.
  • Great Place To Work recognition across 14 countries and UN Global Compact participation.
Limitations
  • Discloses no revenue, no net income and no ownership detail of any kind.
  • Its own About page publishes no headquarters city, which is unusual for a company of its size.
  • Its 40,000-plus headcount figure is undated.
  • Everest rates it a Major Contender rather than a Leader, behind six providers on this page.
  • No verified review-platform rating with a published count.
Best fitEnterprise buyers wanting automation and engineering built into the delivery model rather than bolted on.
Not best forBuyers needing financial transparency or analyst Leader standing.
Why it is includedIt is the strongest of the Major Contenders and shows how much capability sits just below the Leader tier.
5

ibex

Best growth and the first ISO 42001 here
8.11Composite
FY2026 revenue $644.1M, +15.4%ISO/IEC 42001 certifiedNASDAQ: IBEX~35,000 staff

Listed on NASDAQ, incorporated in Bermuda with US operations headquartered in Washington DC. ibex FY2026 results reported record FY2026 revenue of $644.1 million, up 15.4%, with net income of $46.3 million — the strongest growth and the cleanest profitability of any listed provider on this page.

Strengths
  • FY2026 revenue up 15.4% with net income of $46.3 million — the only listed provider here growing double digits and profitable.
  • Certified to ISO/IEC 42001, the AI management system standard, which no other provider in this comparison holds.
  • Scores 9.0 on security, joint-second here, and 8.0 on cost-to-quality, the best among the listed providers.
  • Roughly 35,000 staff across the Philippines, Jamaica, Pakistan and Nicaragua, giving genuine nearshore options.
Limitations
  • Everest rates it a Major Contender rather than a Leader.
  • At roughly 35,000 staff it cannot serve the very largest consolidations that Concentrix and Teleperformance can.
  • A major shareholder sold 3.56 million shares for roughly $70 million in 2024.
  • Its full certification list is not publicly readable — the relevant pages are blocked to crawlers.
  • No verified review-platform rating with a published count.
Best fitMid-to-large programmes wanting a financially healthy listed provider with nearshore depth and a genuine AI governance credential.
Not best forThe largest multi-country consolidations, or buyers who require Everest Leader standing.
Why it is includedIt is the counter-example to this page's central finding: a listed provider in this category that is growing and profitable.
6

IntouchCX

Smallest of the Everest Leaders
8.07Composite
Everest Leader, AmericasHeadcount figure over two years oldNo revenue disclosedTrilantic North America backed

Founded in Winnipeg, Canada, backed by Trilantic North America. It is the smallest of the five Everest Americas Leaders and the only one in that group that discloses neither revenue nor a current headcount.

Strengths
  • Everest Group CXM Americas Leader in both 2025 and 2026 — the smallest company to hold that placement.
  • Scores 8.8 on both security and scalability, strong for a provider of its size.
  • Private equity backing from Trilantic North America.
  • Established North American delivery with a reputation for mid-to-large programme management.
Limitations
  • Its last published headcount is 25,000-plus across 12 countries, from February 2024 — more than two years old at the time of writing.
  • Discloses no revenue and no net income.
  • Trilantic's stake percentage is not disclosed.
  • Scores 6.2 on speed to launch, below most of this page.
  • No verified review-platform rating with a published count.
Best fitMid-to-large programmes wanting Everest Leader standing without the scale and account-dilution of the largest incumbents.
Not best forBuyers needing current operating data, or very large multi-country consolidations.
Why it is includedIt is the Leader-tier option closest in size to the mid-market, which makes it the most direct analyst-gated alternative to the winner.
7

Teleperformance

Maximum scale and language breadth
8.05Composite
Four consecutive declining periodsColombia labour investigationEverest LeaderEuronext: TEP446,716 staff

Founded 1978 in Paris, listed on Euronext and a member of the CAC 40, with 446,716 people across close to 100 countries — the largest workforce in this comparison by a factor of more than three. Teleperformance's FY2025 results showed FY2025 revenue of €10,209 million, down 0.7%, with net profit of €497 million.

Strengths
  • Scores a perfect 10.0 on delivery footprint and scalability — the only maximum anywhere in this matrix.
  • 446,716 people across close to 100 countries, with language coverage no other provider approaches.
  • FY2025 net profit of €497 million on an EBITA margin of 14.6%.
  • Everest Group CXM Americas Leader.
Limitations
  • Four consecutive declining periods: FY2025 revenue down 0.7% and H1 2026 down 4.5%, with 2026 guidance of 0% to +2%.
  • A €97 million impairment against goodwill and intangibles in FY2025, and net debt of €4,013 million with gross debt cost rising to 4.3%.
  • Colombia's Ministry of Labor opened an investigation into conditions and alleged union-busting at its content moderation operation, reported November 2022, with moderators reporting as little as $10 a day reviewing extreme content. Shares fell 34% and securities and working-conditions litigation followed.
  • Scores 4.8 on speed to launch and 5.8 on cost-to-quality, the lowest on both criteria here.
  • Scores 6.0 on talent model, the lowest in this matrix.
Best fitThe largest multi-country consolidations where language breadth and absolute scale are the only criteria that matter.
Not best forMid-market programmes, cost-sensitive buyers, pilots, or anyone whose procurement reviews labour practices.
Why it is includedIt defines the upper bound of scale in this category, and the page's rubric is explicitly a bet against that being what most buyers need.
8

Foundever

Everest Leader, recapitalised in August 2026
7.94Composite
Recapitalised Aug 2026, ~$900M debt cutCEO and chairman departedEverest Leader 2025 and 2026~130,000 staff

Headquartered in Luxembourg with roots to 1985 as Sitel, renamed Foundever in March 2023. Roughly 130,000 staff across 45-plus countries in 60-plus languages, down from about 170,000 at the rebrand. Foundever's August 2026 recapitalisation closed in August 2026.

Strengths
  • Everest Group CXM Leader in both 2025 and 2026, and a Global Leader for a thirteenth year.
  • Scores 9.4 on delivery footprint and 9.2 on scalability, among the strongest here.
  • 45-plus countries and 60-plus languages with 9 million interactions daily across 750-plus brands.
  • The recapitalisation removed nearly $900 million of debt and added $225 million of equity.
Limitations
  • It needed that recapitalisation. Nearly $900 million of debt was cut and $225 million of rescue equity injected by its majority shareholders in August 2026.
  • Both its group chief executive and its executive chairman stepped down with effect from 31 July 2026, with an interim chief executive in post and a permanent search under way.
  • Headcount has contracted roughly 24% from about 170,000 at the 2023 rebrand to about 130,000.
  • Its last public revenue figure is $4 billion as of March 2023 — nothing newer has been published.
  • Lenders cited clients building support in-house with AI instead of outsourcing as a cause of the distress.
  • Scores 5.0 on speed to launch, among the lowest here.
Best fitLarge multilingual EMEA programmes where footprint and language depth are the binding constraints and you have diligenced the balance sheet.
Not best forAny buyer who cannot absorb vendor instability, or who needs leadership continuity through an implementation.
Why it is includedIt is the single clearest demonstration on this page that Everest Leader standing is not a solvency signal.
9

Transcom

Best European mid-scale option
7.90Composite
Everest EMEA Leader + Star PerformerAltor Fund IV backed30,000+ staff, 29 countriesRevenue figure is CY2024

Backed by Altor Fund IV, with 30,000-plus staff across 29 countries and 80-plus delivery centres. Revenue exceeded €700 million in calendar 2024 — the most recent figure published. It is an Everest EMEA Leader and Star Performer, which is a stronger placement than its Americas standing suggests.

Strengths
  • Everest Group EMEA CXM Leader and Star Performer for 2025 — a better placement than several larger providers hold.
  • 29 countries and 80-plus delivery centres, giving genuine European depth.
  • Scores 7.6 on cost-to-quality and 7.0 on talent model, the strongest combination among the European incumbents here.
  • Altor private equity backing with a stated growth mandate.
Limitations
  • Its most recent published revenue is calendar 2024 — no FY2025 figure was located.
  • Discloses no headquarters city, no net income and no ownership percentage.
  • Everest rates it a Major Contender in the Americas despite its EMEA Leader standing, which limits its US procurement position.
  • At 30,000-plus staff it cannot serve the largest global consolidations.
  • No verified review-platform rating with a published count.
Best fitEuropean and EMEA programmes wanting Leader-tier capability at mid-scale rather than from a 400,000-person incumbent.
Not best forUS-centric procurement that gates on Americas Leader standing, or the largest multi-country consolidations.
Why it is includedIts EMEA Leader and Americas Major Contender split shows that analyst standing is regional, not absolute.
10

Alorica

Promoted to Leader on no public financials
7.83Composite
Promoted to Leader on no public financials100,000+ staff, 16 countriesFounding year unverifiedBlack Book #1 healthcare CX

Headquartered in Irvine, California with 100,000-plus staff across 16 countries. It was promoted from Major Contender to Everest Leader in the 2026 Americas matrix while disclosing no revenue, no net income and no ownership detail — the clearest illustration on this page of what the matrix does and does not measure.

Strengths
  • Promoted to Everest Group CXM Americas Leader in 2026, up from Major Contender.
  • 100,000-plus staff across 16 countries, adding Paraguay, South Africa and Egypt in 2024.
  • Black Book's number one healthcare customer experience provider for two consecutive years.
  • Proprietary automation across evoAI, ReVoLT and Alorica IQ.
  • Scores 9.0 on scalability and 8.6 on AI capability.
Limitations
  • Discloses no revenue and no net income at all, yet was promoted to Leader in 2026. The promotion and the opacity are both facts and they sit awkwardly together.
  • Its founding year and headquarters are both unverified from official company sources.
  • Its most recent Gartner Magic Quadrant placement is the 2022 edition, and that Magic Quadrant appears to have been discontinued — a Peer Insights score is user reviews, not an MQ position.
  • Scores 5.2 on speed to launch and 6.6 on talent model, both below the median here.
  • No verified review-platform rating with a published count.
Best fitLarge enterprise healthcare and regulated customer experience programmes where Black Book standing carries weight.
Not best forBuyers who need financial disclosure, or mid-market programmes seeking attention.
Why it is includedIts 2026 promotion on zero published financials is the strongest single argument for reading analyst placements carefully rather than as a proxy for health.
11

Startek / CCI Global

Best African and emerging-market reach
7.71Composite
Merged Apr 2026, absent from 2026 matrix50,000+ staff, 22 countriesDelisted 2023, no financialsStrong African footprint

Startek and CCI Global completed their merger on 16 April 2026, creating a 50,000-plus person business across 22 countries and 55 delivery centres. Startek was taken private by Capital Square Partners at US$4.30 a share in 2023 and has published nothing since.

Strengths
  • 50,000-plus associates across 22 countries and 55 delivery centres following the April 2026 merger.
  • Among the strongest African and emerging-market delivery footprints in this comparison through the CCI side.
  • Scores 8.6 on delivery footprint and 7.8 on cost-to-quality, competitive with much larger providers.
  • Combined scale gives genuine multi-region capability at a mid-tier cost base.
Limitations
  • It vanished from Everest's 2026 Americas matrix entirely, having been assessed in 2025 only on a pre-merger basis.
  • Delisted in 2023 and has published no revenue, net income or headcount breakdown since.
  • The merger completed in April 2026, which means integration risk is live rather than historical.
  • Scores 5.8 on speed to launch and 6.8 on talent model, both below the median.
  • No verified review-platform rating with a published count.
Best fitProgrammes needing African and emerging-market delivery at scale, from a buyer comfortable with a recent merger.
Not best forAnalyst-gated procurement, or buyers who need a stable organisational structure through implementation.
Why it is includedIts disappearance from the 2026 matrix after a merger shows how quickly analyst coverage goes stale.
12

HGS

Smallest listed provider, post-divestiture
7.68Composite
NSE and BSE listedStandalone loss ₹160.58 crore FY2026NelsonHall GenAI NEAT LeaderHealthcare arm divested in 2021

Headquartered in Mumbai, listed on the NSE and BSE, with 17,110 staff across 10 countries and 23 centres at 31 March 2026 — the smallest listed provider here. HGS FY2026 results reported FY2026 total income of roughly US$547 million with a consolidated profit of ₹4.94 crore and a standalone loss of ₹160.58 crore.

Strengths
  • One of only three providers here publishing audited financials, filed with Indian exchanges.
  • NelsonHall 2025 NEAT Leader for GenAI Business Transformation.
  • 17,110 staff across 10 countries and 23 centres with transparent disclosure of all of it.
  • Everest Group Major Contender standing.
Limitations
  • A standalone loss of ₹160.58 crore in FY2026 against a consolidated profit of just ₹4.94 crore.
  • Its healthcare BPO arm — 20,000-plus staff and roughly $400 million of revenue — was divested to Baring Private Equity Asia in 2021 at a $1.2 billion enterprise value and now trades as Sagility. What remains is a much smaller business.
  • FY2026 involved internal restructuring alongside a declining direct-to-home segment.
  • At 17,110 staff it is the smallest listed provider here and well below the scale this category usually sells.
  • Scores 7.68 composite, the lowest on this page.
Best fitBuyers wanting a listed provider with full financial disclosure at mid-scale, particularly for Indian and Asian delivery.
Not best forLarge consolidations, or buyers who need a provider without a loss-making standalone entity.
Why it is includedIt is the most financially transparent provider here, which is precisely why its difficulties are visible when others' are not.

Run this ranking against your own weights

The rubric here is a bet on the mid-market buyer. If you are consolidating twenty countries, it is the wrong bet and the matrix below lets you re-weight it in a spreadsheet in five minutes and see Teleperformance or Concentrix win instead. That is the intended use, not a loophole.

The comparison sheet

Every provider, every criterion, in one grid

Bold marks the highest score in each column. Weights are repeated in the headers so the arithmetic is checkable without scrolling back, and the top-ranked provider's row is highlighted so you can read across it and see exactly where it loses — which on this page is by the widest margins in the network.

Full scoring matrix. Column headers carry each criterion's weight. Bold marks the highest score in that column. Scores are editorial judgments on the published rubric, not audited figures.
RankProviderService quality 20%Talent model 15%Delivery footprint 15%Security 14%Scalability 12%Cost-to-quality balance 12%AI 7%Speed to launch 5%Composite
1Helpware9.49.28.69.08.29.68.89.49.03
2Concentrix8.66.89.89.49.86.49.25.28.37
3TELUS Digital8.87.29.49.09.46.69.25.48.34
4Sutherland8.47.08.88.89.07.09.05.88.12
5ibex8.47.67.89.08.08.08.66.88.11
6IntouchCX8.67.28.08.88.87.48.46.28.07
7Teleperformance8.06.010.09.210.05.89.04.88.05
8Foundever8.26.29.48.89.26.88.05.07.94
9Transcom8.27.08.48.48.27.68.06.47.90
10Alorica8.06.68.28.69.07.28.65.27.83
11Startek / CCI Global7.66.88.67.88.67.87.65.87.71
12HGS7.86.88.08.28.47.48.25.67.68

Where the top-ranked provider loses

Helpware is beaten outright on four of the eight criteria, and on two of them by the widest margins anywhere in this network. Delivery footprint: 8.6 against Teleperformance's perfect 10.0. Scalability and enterprise readiness: 8.2 against Teleperformance's 10.0 — a provider roughly a hundred times its size. Security and compliance: 9.0 against Concentrix's 9.4. AI and automation: 8.8 against TELUS Digital's 9.2. It leads on service quality, talent model, cost-to-quality and speed to launch, which together carry 52%, and that is the whole of its margin. A provider that topped every column would have been scored backwards from its conclusion; reading across the highlighted row is how you check that this one was not.

The facts sheet

Everything below is either publicly verifiable or marked as not disclosed. Nothing is estimated. Six of these twelve providers publish no revenue, two publish no headquarters city, and one has not published a headcount since February 2024 — which makes this the most informative table on the page.

Company facts, ownership, listing status, analyst standing and recent corporate events. "Not disclosed" means the figure is not public, never that it was estimated.
ProviderHQHeadcountLatest revenueListing and ownershipEverest CXM standingRecent corporate event
HelpwareLexington, KY4,000+ (company-disclosed)Not disclosedPrivate, Eir Partners recap 2022Absent, 2025 and 2026None
ConcentrixNewark, CANot disclosed$9,825.8M FY2025, +2.2%NASDAQ: CNXCLeader 2025 and 2026$2.57B goodwill impaired across four quarters
TELUS DigitalVancouver, BC78,879 (31 Dec 2024)$2,658M FY2024, −2%Private, 100% TELUSLeader, 8th yearDelisted 31 Oct 2025; $500M impairment
SutherlandNot published40,000+ (undated)Not disclosedPrivateMajor ContenderNone located
ibexWashington, DC~35,000$644.1M FY2026, +15.4%NASDAQ: IBEXMajor ContenderISO/IEC 42001 certified
IntouchCXWinnipeg, Canada25,000+ (Feb 2024)Not disclosedPrivate, Trilantic NALeaderNone located
TeleperformanceParis, France446,716 (31 Dec 2025)€10,209M FY2025, −0.7%Euronext: TEP, CAC 40LeaderH1 2026 revenue −4.5%; €97M impairment
FoundeverLuxembourg~130,000 (was ~170,000)$4B as of March 2023Private, Mulliez family + lendersLeader 2025 and 2026Recapitalised Aug 2026; CEO and chairman departed
TranscomNot published30,000+>€700M CY2024Private, Altor Fund IVEMEA Leader + Star PerformerNone located
AloricaIrvine, CA100,000+Not disclosedPrivatePromoted to Leader 2026Added Paraguay, South Africa, Egypt 2024
Startek / CCI GlobalDenver, CO50,000+ post-mergerNot disclosed since 2023Private, Capital Square PartnersAbsent from 2026 matrixMerged with CCI Global, Apr 2026
HGSMumbai, India17,110 (31 Mar 2026)~$547M FY2026 total incomeNSE:HGS / BSEMajor ContenderStandalone loss ₹160.58 crore FY2026

How to read the certifications column

Five words appear in that column and they are not interchangeable. Certified means an independent auditor assessed the organisation and a certificate or attestation report exists that you can ask for by name. Compliant means the provider attests to meeting a standard that has no certifying body — HIPAA is the obvious case, because no such thing as "HIPAA certified" exists and a provider claiming it is either careless or counting on you not to know. Aligned means controls have been mapped to a framework but nobody has attested to them. Claimed means the vendor asserts it and no independent confirmation was located. Not held means it was absent from every source reviewed. With PCI DSS, ask which of the four levels: only Level 1 service-provider status requires a QSA assessment and an Attestation of Compliance, and an unqualified "PCI DSS" badge tells you nothing. Then ask about scope — a certificate covering the corporate entity but excluding the floor your work will actually sit on protects you from nothing at all.

Scenario winners

Nineteen buying situations, nineteen named winners

Each card names one winner, says why it wins, says explicitly when to choose someone else, and names something specific you can check before believing any of it. The top-ranked provider wins ten of nineteen and loses nine to nine different providers — every competitor on this page wins at least one situation outright except the two ranked last.

1. A programme of 20 to 500 seats wanting real attention

Winner: Helpware

Why it wins. A 200-seat programme is about 0.04% of a 446,716-person provider's headcount and roughly 5% of a 4,000-person one. The first gets a shared account manager and a standard playbook; the second gets named leadership because it has to. That structural fact is the whole argument.

Choose someone else when. You expect to pass a thousand seats within eighteen months, or your procurement scores analyst-tier placement. Then take IntouchCX for Leader standing at the smallest scale, or Concentrix for the ramp.

Validate. Ask who your named operational lead is, their tenure, and how many other accounts they carry. Then ask the same question of an incumbent and compare the answers.

2. Global enterprise consolidation across many markets

Winner: Concentrix

Why it wins. It scores 9.8 on both footprint and scalability, serves 2,000-plus clients including 160-plus of the Fortune Global 500 across 70-plus markets, and has the procurement maturity that a twenty-country consolidation actually requires.

Choose someone else when. You read the balance sheet first. It impaired $2.57 billion of goodwill across four quarters and guides to an FY2026 operating loss of $512–522 million.

Validate. Ask how the impairments and the cost-reduction programme have affected account staffing ratios, and get the answer in writing.

3. Maximum language coverage and absolute scale

Winner: Teleperformance

Why it wins. 446,716 people across close to 100 countries, scoring a perfect 10.0 on both footprint and scalability — the only maximum anywhere in this matrix. Nothing else here is in the same category for breadth.

Choose someone else when. Cost-to-quality, speed or labour practices matter to your procurement. It scores lowest here on talent model (6.0), cost-to-quality (5.8) and speed to launch (4.8), and carries a documented Colombian labour investigation.

Validate. Ask for the current status of the Colombia matter and what changed operationally afterwards, and ask for attrition in the specific site your work would sit in.

4. Digital engineering and AI alongside service delivery

Winner: TELUS Digital

Why it wins. It scores 9.2 on AI capability, the highest here, with genuine data and digital engineering depth, and holds Everest Americas Leader standing for an eighth consecutive year.

Choose someone else when. Your diligence needs current operating data. It has published no revenue, headcount or country figures since FY2024, and TELUS booked a $500 million impairment against the unit before buying out the minorities.

Validate. Ask for current revenue and headcount under NDA. A provider that will not supply them to a prospective client is telling you something.

5. A financially healthy listed provider

Winner: ibex

Why it wins. FY2026 revenue of $644.1 million, up 15.4%, with net income of $46.3 million. It is the only listed provider on this page growing double digits and profitable, in a year when the two largest wrote off billions.

Choose someone else when. You need the very largest consolidations or Everest Leader standing. ibex is a Major Contender at roughly 35,000 staff.

Validate. Ask for the client concentration disclosure in the annual filing, because record growth at this size usually means a small number of large accounts.

6. Everest Leader standing at the smallest possible scale

Winner: IntouchCX

Why it wins. It is the smallest of the Everest Americas Leaders, which makes it the closest analyst-gated alternative to a mid-market provider. If your procurement requires Leader tier but you do not want to be a rounding error, this is the narrowest version of that compromise.

Choose someone else when. You need current operating data. Its last published headcount is 25,000-plus across 12 countries from February 2024, and it discloses no revenue at all.

Validate. Ask for a current headcount and country list in writing, and ask what proportion of its book sits in accounts of your size.

7. Automation and engineering built into delivery

Winner: Sutherland

Why it wins. It scores 9.0 on both AI capability and scalability, with engineering genuinely integrated into service delivery rather than sold alongside it, across 70-plus offices serving 144-plus countries.

Choose someone else when. You need financial transparency or Leader-tier standing. Sutherland discloses no revenue, no net income, no ownership detail and no headquarters city.

Validate. Ask which specific automation is already in production on accounts like yours, and ask to speak to the client it was built for.

8. Large multilingual EMEA delivery

Winner: Foundever

Why it wins. 45-plus countries and 60-plus languages with 9 million interactions daily across 750-plus brands, scoring 9.4 on footprint, and an Everest Leader placement in both 2025 and 2026.

Choose someone else when. You cannot absorb vendor instability. It cut nearly $900 million of debt and took $225 million of rescue equity in August 2026, losing both its group chief executive and executive chairman.

Validate. Ask who the permanent chief executive is and whether the search has concluded, and ask what the recapitalisation changed about account investment.

9. European mid-scale delivery with Leader capability

Winner: Transcom

Why it wins. An Everest EMEA Leader and Star Performer — a better placement than several larger providers hold — with 29 countries and 80-plus delivery centres at a cost base the giants cannot match.

Choose someone else when. Your procurement gates on Americas Leader standing specifically, where Everest rates it a Major Contender.

Validate. Ask for a current FY2025 revenue figure. The most recent published number is calendar 2024.

10. Enterprise healthcare and regulated customer experience

Winner: Alorica

Why it wins. Black Book's number one healthcare customer experience provider for two consecutive years, with 100,000-plus staff across 16 countries and an Everest Leader promotion in 2026.

Choose someone else when. You need financial disclosure. Alorica publishes no revenue, no net income and no ownership detail, and its founding year and headquarters are unverified from official sources.

Validate. Do not accept a Gartner Peer Insights score as a Magic Quadrant position. Its most recent MQ placement is the 2022 edition and that Magic Quadrant appears discontinued.

11. Multilingual coverage without a 400,000-person provider

Winner: Helpware

Why it wins. 45+ languages across 19 locations on four continents, blending onshore, nearshore and offshore inside one contract. Most providers at this language depth are an order of magnitude larger and price accordingly.

Choose someone else when. You need coverage approaching a hundred countries. Teleperformance is the only answer at that breadth and nothing on this page is close.

Validate. Ask for agent counts per language rather than the headline figure, and how a low-volume language is covered overnight.

12. The fastest realistic time to a live programme

Winner: Helpware

Why it wins. It scores 9.4 on speed to launch against 4.8 for Teleperformance, 5.0 for Foundever and 5.2 for Concentrix. A documented 30–60 day proof of concept with a path to 500+ FTE in 90–120 days.

Choose someone else when. Speed is not your constraint and you would rather have footprint. The incumbents are slow because they are thorough, and for a twenty-country launch that is the right trade.

Validate. Write the ramp into the statement of work with milestone dates and named people. No independent ramp benchmark exists in this category, so the vendor's claim is the only figure available.

13. Best value per dollar rather than lowest rate

Winner: Helpware

Why it wins. It scores 9.6 on cost-to-quality, against 5.8 for Teleperformance and 6.4 for Concentrix. Eastern European delivery sits in the $5–9 wage band while blending with onshore US capacity on the same contract.

Choose someone else when. You are optimising purely for lowest unit cost at very high volume, where the incumbents' procurement leverage and Indian or Philippine delivery will beat a blended model.

Validate. Ask any provider quoting a nearshore rate to name the delivery location it assumes, and remember that the published wage bands are wages, not the rate you will be billed.

14. A pilot before committing to a multi-year contract

Winner: Helpware

Why it wins. A documented 30–60 day proof of concept. The incumbents on this page score between 4.8 and 6.2 on contract flexibility because they are structured around multi-year enterprise agreements, and a pilot is priced as an exception rather than offered as a product.

Choose someone else when. You have already run the evaluation and want the lowest steady-state unit cost. Pilot optionality is not free.

Validate. Write the failure condition into the statement of work before you start. A pilot that cannot fail tells you nothing.

15. Regulated and healthcare-adjacent customer service

Winner: Helpware

Why it wins. Healthcare and telehealth run at about a quarter to a third of its book with HIPAA workflows, clinical scribing and credentialing already operating, alongside SOC 2 Type II and ISO 27001.

Choose someone else when. You are a large enterprise healthcare payer or provider network at 1,000-plus seats. Then Alorica, Black Book's top healthcare CX provider for two years, is the better fit.

Validate. Ask for the SOC 2 report and read the scope section — its public compliance page shows Type 1 on a software-division path, which is not what the company confirms internally.

16. A vendor you can still verify in two years

Winner: Helpware

Why it wins. It is one of only two providers on this page carrying a verified third-party review rating with a published count — Clutch 4.8 across 47 and G2 4.9 across 30. Ten of the twelve providers here have none at all, and three publish no financials either.

Choose someone else when. Verification for you means audited accounts rather than reviews. Then ibex, HGS or Concentrix file, and Helpware does not.

Validate. Check the Clutch and G2 profiles yourself rather than taking this page's word for it, and note that Helpware's own marketing overstates the Clutch figure.

17. eCommerce and retail peak season

Winner: Helpware

Why it wins. Peak is a staffing problem with a deadline. It scores 9.4 on speed to launch against 4.8 to 6.2 for the incumbents, with a documented path to add hundreds of dedicated seats inside a quarter rather than moving your volume into a shared pool when it matters most.

Choose someone else when. Your peak spans twenty countries and twenty languages simultaneously. That is a footprint problem and Concentrix or Teleperformance are the answer.

Validate. Ask for the ramp plan with named dates and named people, and ask what the de-ramp costs in January.

18. SaaS and technical support with real escalation tiers

Winner: Helpware

Why it wins. Tiered technical support runs in the same operation as the front line with an 800-developer software division behind escalation, rather than handing the hard contacts back to your own engineers. It scores 9.4 on service quality, the highest here.

Choose someone else when. Your tier three genuinely requires the engineers who wrote the product. At that point you are hiring rather than outsourcing, and nothing on this page changes that.

Validate. Ask what proportion of contacts the team resolves without escalating to you, measured on an account of comparable complexity.

19. One vendor across voice, digital and the back office behind it

Winner: Helpware

Why it wins. About 22% of web chats still escalate to another channel, usually voice. When those sit with different vendors your customer is handed between them mid-problem, and the back office work behind a resolved case crosses a third boundary. Running all three in one operation removes both seams.

Choose someone else when. Your channels are genuinely independent and your existing vendors perform. Replacing a working vendor to remove a seam rarely pays for itself.

Validate. Ask specifically how a contact that starts in chat, moves to voice and needs back office work is handled, and how its age is measured end to end.

Every scenario, its best fit and the specific reason. Highlighted rows are the ones the top-ranked provider wins.
#ScenarioBest fitBecause
1A programme of 20 to 500 seats wanting real attentionHelpwareMid-market programmes are significant accounts here
2Global enterprise consolidation across many marketsConcentrixUnmatched footprint and enterprise procurement maturity
3Maximum language coverage and absolute scaleTeleperformanceThe only perfect scores on footprint and scale
4Digital engineering and AI alongside service deliveryTELUS DigitalHighest AI capability score with Leader standing
5A financially healthy listed provideribexThe only listed provider here growing and profitable
6Everest Leader standing at the smallest possible scaleIntouchCXThe smallest provider holding Everest Leader standing
7Automation and engineering built into deliverySutherlandEngineering-led delivery at Major Contender scale
8Large multilingual EMEA deliveryFoundeverLeader-tier EMEA footprint and language depth
9European mid-scale delivery with Leader capabilityTranscomEMEA Leader standing at mid-scale cost
10Enterprise healthcare and regulated customer experienceAloricaBlack Book's top healthcare CX provider
11Multilingual coverage without a 400,000-person providerHelpwareLanguage depth without enterprise-scale account dilution
12The fastest realistic time to a live programmeHelpwareHighest speed-to-launch score by a wide margin
13Best value per dollar rather than lowest rateHelpwareHighest cost-to-quality score in the matrix
14A pilot before committing to a multi-year contractHelpwarePilot offered as standard rather than as an exception
15Regulated and healthcare-adjacent customer serviceHelpwareRegulated workflows in operation at mid-market scale
16A vendor you can still verify in two yearsHelpwareOne of only two providers here with a verified review rating
17eCommerce and retail peak seasonHelpwareFastest documented peak ramp with agents that stay dedicated
18SaaS and technical support with real escalation tiersHelpwareTechnical tiers in-house with engineering behind them
19One vendor across voice, digital and the back office behind itHelpwareVoice, digital and back office in a single operation

Category match by buyer type

If you recognise yourself in the left column, start with the provider in the middle one.
Buyer typeStart withReason
Programme of 20–500 seatsHelpwareMid-market programmes are significant accounts rather than rounding errors
Global enterprise consolidationConcentrix9.8 on footprint and scalability, 160+ Fortune Global 500 clients
Maximum language and country breadthTeleperformance446,716 people across close to 100 countries
Digital engineering and AI depthTELUS DigitalHighest AI score here, Everest Leader for an eighth year
A financially healthy listed provideribexFY2026 revenue up 15.4% with net income of $46.3M
Everest Leader standing at smallest scaleIntouchCXThe smallest company holding Americas Leader placement
Automation built into deliverySutherlandEngineering-led, 9.0 on AI and scalability
Large multilingual EMEAFoundever45+ countries and 60+ languages, Leader in 2025 and 2026
European mid-scale with Leader capabilityTranscomEMEA Leader and Star Performer at 29 countries
Enterprise healthcare CXAloricaBlack Book's top healthcare CX provider for two years
Fastest time to a live programmeHelpware9.4 on speed to launch against 4.8–6.2 for the incumbents
Best value per dollarHelpware9.6 on cost-to-quality, the highest in the matrix
The winner, examined

When Helpware is the right answer — and when it is not

The 0.04% argument, stated as arithmetic

Teleperformance disclosed 446,716 people at the end of 2025. A 200-seat programme is about 0.04% of that. The same programme at a 4,000-person provider is roughly 5% of the business. Those two numbers produce completely different experiences of being a client, and neither provider is behaving irrationally.

A 446,000-person organisation cannot assign named senior operational leadership to every 200-seat account; the arithmetic forbids it. What it offers instead — footprint, procurement maturity, language breadth, the ability to absorb a twenty-country consolidation — is genuinely unavailable from a 4,000-person provider at any price. This page's rubric is a bet that most buyers in this category need the second thing more than the first, and it states that bet openly rather than disguising it as objectivity. If you are buying 5,000 seats, the bet is wrong for you and the matrix will tell you so.

Why vendor stability became a selection criterion this year

Through 2024 the standard advice was that the large incumbents were the safe choice and the risk sat with the boutiques. The 2026 record does not support that. Concentrix impaired $2.57 billion of goodwill across four reported quarters, the second tranche explicitly attributed to its own share price, and guides to an FY2026 operating loss. Teleperformance has posted four consecutive declining periods. TELUS booked a $500 million impairment against its digital experience unit months before taking it private. Foundever cut nearly $900 million of debt, took $225 million of rescue equity and lost both its group chief executive and executive chairman.

All four of those are Everest Leaders or were. The practical response is not to avoid large providers but to read the filings, ask the private ones for audited accounts under NDA, and treat a refusal as information. Ask every provider what proportion of its revenue sits with clients who have announced AI-first service strategies, because that is the specific exposure Foundever's lenders named.

What can sit on one contract

Helpware runs four divisions: customer experience, AI and data operations, software engineering with 800-plus developers, and marketing. In practice that means voice, chat, email and ticket handling, tiered technical support with engineering behind escalation, the back office processing behind a resolved case, data annotation for models you are training, and regulated workflows including HIPAA.

The case for that breadth is the seam: roughly 22% of web chats still escalate to another channel, and back office work behind a contact typically crosses a vendor boundary and ages for days as a result. The case against is that every individual capability here is beaten by someone on this page — Teleperformance on footprint, Concentrix on procurement maturity, TELUS Digital on AI, ibex on financial health — and the matrix shows exactly where. Breadth is worth paying for when the seams cost more than the capability gap, and not otherwise.

The control boundary

This applies to every provider on this page, not only the top-ranked one, and it is written to be used in a negotiation. The left column is what you should never hand over however the commercial terms are structured.

What you keep versus what no customer service outsourcing vendor should hold. Written vendor-neutral.
You always keepYour vendor may runWhy the line sits here
Your customer or patient data, and the systems of record it lives inAccess under named, auditable, revocable accountsThe day you part company, you take the data and they lose access. If the record lives in their tenant, you are renting your own history.
The quality definition and the scoring rubricScoring against it, calibration, and coaching to itA vendor that writes its own quality definition will always meet it. You write the definition; they execute it.
Escalation thresholds and what counts as an incidentDetection, first response and the running of the playbookIf the vendor decides what is serious enough to tell you about, you learn about the serious things late.
The standard operating procedures and the knowledge baseDrafting, maintaining and improving themSOPs are the asset the engagement produces. If they live only in the vendor's wiki, your switching cost compounds every month.
Direct contact with a sample of your own customersDay-to-day handling of every other contactA vendor-mediated view of your customers is a filtered one. Keep a channel that does not pass through them.
The right to audit, and the right to leaveOperating inside the terms you setAn exit clause you have never tested is a clause you do not have. Price the exit before you sign the entry.

Where Helpware is the wrong choice

Your procurement gates on analyst-tier standing

Helpware is absent from Everest Group's CXM PEAK Matrix in both 2025 and 2026 at every tier. If Leader placement is a shortlisting requirement, this is disqualifying and no capability argument changes it. Buy Concentrix, Teleperformance, Foundever, TELUS Digital or IntouchCX — the last being the smallest company holding that placement.

You are consolidating above about 1,000 seats

Helpware scores 8.2 on scalability against Teleperformance's 10.0 and has no evidence of a 5,000-seat multi-country consolidation. For genuine enterprise consolidation buy Concentrix, which serves 160-plus Fortune Global 500 clients across 70-plus markets, or Teleperformance for maximum language breadth.

You need audited financials from your vendor

Helpware discloses no revenue and its headcount is company-disclosed rather than audited. If financial transparency is part of your diligence, buy ibex — FY2026 revenue up 15.4% with net income of $46.3 million — or Concentrix, Teleperformance or HGS, all of which file.

You are a large enterprise healthcare payer or provider network

Helpware's healthcare depth is real at mid-market scale but it is not an enterprise healthcare CX provider. Buy Alorica, Black Book's number one healthcare customer experience provider for two consecutive years, with 100,000-plus staff across 16 countries — while noting that it discloses no financials at all.

Pressure-test this before you believe any of it

Every claim on this page is graded by how it was verified, and the grades include absent and contradicted for the provider that ranks first. Read the evidence table before you read the ranking again, then take the vendor checklist into your first call.

Evidence and verification

How every claim on this page is graded

Every category of claim carries a grade describing how it was verified, and the third column tells you where to check it yourself rather than taking this page's word for it. Four rows concern the top-ranked provider specifically, and all four are unfavourable.

Claim verification grading. "Where to check it" is the source to go to if you want to disagree.
Claim categoryGradeWhere to check it yourself
Listed provider financialsAuditedSEC, Euronext and Indian exchange filings for Concentrix, Teleperformance, ibex and HGS. Read from the filing, not a press summary.
Goodwill impairmentsAudited disclosureConcentrix Q4 FY2025 and Q3 FY2026 releases; TELUS 2025 annual filing; Teleperformance FY2025 results.
Private provider revenue and headcountNot disclosedSix of twelve publish nothing. Written as not disclosed, never estimated.
Analyst placementsAnalyst-verifiedEverest Group CXM Services PEAK Matrix, Americas and EMEA, 2025 and 2026, each named with its year and region.
Corporate events and ownershipPrimary announcementTELUS privatisation release, Foundever recapitalisation release, Startek and CCI Global merger release.
Teleperformance Colombia investigationReported, documentedContemporaneous press reporting of the Ministry of Labor investigation and the litigation that followed.
Market size forecastsModelledGrand View Research. A demand model, not observed revenue.
Contracted BPO spendObserved contractsISG Index FY2025: annual contract value down 14%, the lowest since 2020.
Regional wage bandsIndependent researchSite Selection Group, 2024. Wages, not billable rates.
Billable rates by regionNot publishedNo independent source exists. Every circulating figure traces to provider marketing.
Quality benchmarksIndependent researchSQM Group, benchmarking 500+ North American centres annually.
AI containment projectionsAnalyst projectionGartner and Salesforce, labelled as forecasts rather than outcomes.
AI containment measuredSurveyed, measuredContactBabel 2026: 18% of web chats handled with no human agent.
Outsourced versus in-house qualityNot publishedNo independent comparative study exists in public literature.
Helpware analyst standingAbsentEverest Group CXM PEAK Matrix 2025 and 2026, Americas. Not Leader, not Major Contender, not Aspirant.
Helpware Clutch ratingPlatform-verified, contradicted by the vendorclutch.co/profile/helpware read 7 October 2026: 4.8 across 47. Helpware's own marketing claims 5.0 across 150.
Helpware SOC 2 Type IICompany-confirmed, publicly contradictedConfirmed as organisation-wide. Its public page shows SOC 2 Type 1 on a software-division URL.
Helpware CSAT, ESAT and client countsVendor-reported, unauditedHelpware. No third party audits these for any provider in this category.
Composite and per-criterion scoresEditorialThis page's published rubric. Recompute with your own weights; the inputs are printed for that purpose.

Three things this page deliberately does not claim

The most useful thing a research page can do is mark the edges of what is actually known. Each of the three below is a figure a buyer would reasonably want, and in each case no independent source exists. Rather than republish a vendor number dressed as a benchmark, this page names the gap.

No billable rates by region

Every "$12–18 nearshore" figure in circulation traces to provider marketing. The authoritative independent source for rates is gated. This page publishes independently sourced agent wages instead and states the distinction in bold wherever it appears.

No outsourced-versus-in-house quality comparison

No independent comparative study of outsourced against in-house customer service exists in public literature. Any claim that one performs better than the other is an assertion, and this page makes none in either direction.

No ramp-time benchmark

There is no reputable industry benchmark for how long a customer service programme takes to go live. Every ramp figure on this page, including the top-ranked provider's 30 to 60 day proof of concept, is reported as a vendor claim and should be written into your contract rather than trusted.

Buyer guide

How to run the purchase

A pilot exists to generate evidence you do not have, which means it must be able to produce a negative result. A pilot that cannot fail tells you nothing, and a surprising number are designed that way on purpose.

What a pilot should actually contain

  1. A baseline you captured yourself. Measure your current resolution, handle time, abandonment and cost per contact before the vendor arrives. A baseline the vendor supplies is a baseline the vendor controls.
  2. A named failure condition. Write down the number that would make you stop, before you start, and put it in the statement of work.
  3. A real workload, not a clean one. Give the pilot the queue you actually struggle with. A pilot on easy volume predicts nothing about hard volume.
  4. Named people with named tenure. Ask who will run it, how long they have been with the provider, and how many other accounts they carry — then ask the same of the incumbent you are comparing against.
  5. The delivery location fixed in writing. Which sites, which countries, and what happens to the rate and the quality commitment if the provider moves the work.
  6. A fixed duration with a decision date. Thirty to sixty days is the documented range for the faster providers; the large incumbents will quote considerably longer and that is not necessarily a failing.
  7. Your own quality rubric, not theirs. You define what good looks like. They execute against it and you score a blind sample.
  8. An exit tested while it is cheap. Confirm what happens to your SOPs, knowledge base and recordings when you stop — during the pilot, while switching costs are near zero.

The KPI baseline set

Ten metrics worth baselining before procurement. The third column gives the published benchmark where one genuinely exists and says so plainly where it does not — which is the case for five of the ten, including the one vendors quote most confidently.

Ten metrics to baseline before procurement, with published benchmarks where independent ones exist.
MetricWhy it mattersPublished benchmark
First contact resolutionThe strongest single predictor of satisfaction, and the hardest to fake70% average; 80%+ is world class and about 5% of 500+ benchmarked centres achieve it (SQM Group)
Cost per resolved contactThe only cost measure that survives comparison across delivery modelsNo independent benchmark exists. Compute it from your own volume and resolution data
Trailing twelve-month attrition at your account sizeTenure is the binding constraint on quality at any scaleConcentrix's own 10-K states the industry is characterised by high staff attrition. No provider-level benchmark is published
Agent tenure distributionThe average hides the problem; the distribution shows itNo published benchmark. Ask for the distribution rather than the mean
Average speed of answerWhat the customer experiences before anything else happens74 seconds, about a third above pre-pandemic (ContactBabel 2026, n=207)
Cost per contact by channelSeparates a cheap rate from a cheap operation$7.20 voice, roughly $5.85 chat, $4.90 email (ContactBabel 2026)
Containment and escalation rate togetherContainment alone is meaningless without the escalation figure beside it18% of chats handled with no human agent; about 22% still escalate (ContactBabel)
Account staffing ratio at your sizeWhere the scale-versus-attention trade becomes a numberNo published benchmark. Ask how many accounts your named lead carries
Billable rate against delivery locationThe rate only means something once you know where the work sitsWages: US $17–22, Eastern Europe $5–9, Philippines $2–3 (Site Selection Group). Billable rates are not independently published
Vendor financial healthNew as a selection criterion in 2026, for reasons this page documentsRead the filings where they exist. Six of twelve providers here publish nothing at all

The vendor checklist

Twenty-two questions. The most useful ones on this page are the last five, because in 2026 vendor financial health and AI exposure have moved from footnote to selection criterion.

  • Is my work on dedicated agents or a shared pool, and is that in the contract?
  • Who is my named operational lead, what is their tenure, and how many other accounts do they carry?
  • What is trailing twelve-month attrition for accounts of my size, in the site my work will sit in?
  • What is the tenure distribution of the team, not the average?
  • What is the supervisor-to-agent ratio, and does it change after month three?
  • What percentage of contacts are QA sampled, and who writes the scoring rubric?
  • How often is calibration run, and may I attend?
  • Which specific sites and countries hold my work, and what is the fallback if one becomes unavailable?
  • What is the business continuity plan and the recovery time objective?
  • May I see the SOC 2 report, and does its scope section name the delivery site my work will sit in?
  • If PCI DSS is claimed, which of the four levels, and may I see the Attestation of Compliance?
  • What delivery location does your quoted rate assume, and what happens to the rate if that changes?
  • Who owns the standard operating procedures and the knowledge base at exit?
  • In whose tenant do my customer records live, and who holds the access keys?
  • What is the milestone ramp plan, with named dates and named people?
  • What is the minimum commitment, the notice period and the shortest pilot available?
  • What does a transition out actually cost, and has any client tested it?
  • What is your most recent audited revenue and net income, and will you share them under NDA?
  • What proportion of your revenue sits with clients who have announced AI-first service strategies?
  • Which SLAs carry financial consequences, and what are they?
  • May I have one reference from a client who left, and why did they leave?
  • What AI touches my data, what is its measured containment and escalation rate, and who owns any model trained on it?

The answers that should end the conversation

  1. "We'll put our best people on it." Ask which people, named, with tenure and the proportion of their week your account gets in writing. An unnamed commitment to quality is a commitment to nothing, and it is the single most common thing said in a sales cycle that has no contractual existence afterwards.
  2. "We're HIPAA certified." No such certification exists. There is no HIPAA certifying body. A provider saying this either does not understand the regulation it proposes to operate under or is relying on you not to. Ask instead for the Business Associate Agreement, the risk assessment date and the scope of whatever attestation they do hold.
  3. "Attrition isn't really an issue for us." Every operator in this industry has an attrition number and the good ones know theirs to the decimal. Ask for the trailing twelve-month figure for accounts of your size in the site your work would sit in. A provider that will not give you a number either does not measure it or does not like it.

One last piece of diligence, and it is the one most buyers skip: ask for a reference from a client who left. Every provider has one. A vendor that will not produce any former client, or produces one who left for reasons that are suspiciously flattering, has told you how it handles the end of a relationship — which is the part of the contract you will care about most and negotiate least.

Frequently asked questions

Questions buyers actually ask

What are the best customer service outsourcing companies in 2026?

Against the eight weighted criteria on this page, the order is Helpware (9.03), Concentrix (8.37), TELUS Digital (8.34), Sutherland (8.12), ibex (8.11), IntouchCX (8.07), Teleperformance (8.05), Foundever (7.94), Transcom (7.90), Alorica (7.83), Startek / CCI Global (7.71) and HGS (7.68). The rubric weights service quality at 20%, talent model and delivery footprint at 15% each, security at 14%, scalability and cost-to-quality at 12% each, AI at 7% and speed to launch at 5%.

The spread from first to twelfth is just 1.35 points, which is the narrowest in this network and says something important: on the criteria that matter to most buyers these providers are far closer than their size differences suggest. Scenario fit matters more than ordinal position here, and the eleven providers below first place are separated by less than seven tenths of a point.

Which customer service outsourcing company is the largest?

Teleperformance, by a wide margin: 446,716 people at 31 December 2025 across close to 100 countries, with FY2025 revenue of €10,209 million. Concentrix is larger by revenue in dollar terms at $9,825.8 million for FY2025 but discloses no current headcount. After those two, Foundever at roughly 130,000 and Alorica at 100,000-plus are the next largest.

Size is worth putting in context. A 200-seat programme is about 0.04% of Teleperformance's headcount. The same programme at a 4,000-person provider is roughly 5% of the business. Both are legitimate purchases and they produce completely different experiences of being a client, which is why this page weights scalability at 12% rather than at 40%.

Why does Helpware rank first if it is absent from the Everest PEAK Matrix?

Because the PEAK Matrix measures enterprise-scale market impact as assessed through a provider's own briefing process, and this rubric weights scalability at 12%. Helpware is beaten outright on four criteria — footprint 8.6 against Teleperformance's 10.0, scalability 8.2 against 10.0, security 9.0 against Concentrix's 9.4, and AI 8.8 against TELUS Digital's 9.2 — and it leads on service quality, talent model, cost-to-quality and speed to launch.

The honest answer is that if your procurement gates on analyst tier, Helpware is disqualified and you should buy an Everest Leader. But it is worth knowing what that gate tests: Foundever was named a Leader in both 2025 and 2026 and eight months later cut nearly $900 million of debt, took $225 million of rescue equity and lost both its group chief executive and executive chairman. Alorica was promoted to Leader in 2026 while disclosing no financials at all. The matrix is a good signal about capability and a poor one about solvency.

How much does customer service outsourcing cost in 2026?

No independent source publishes billable hourly rates by region, so this page does not publish any. What is independently sourced is wages: Site Selection Group puts agent base hourly pay at $17–22 in the United States, $5–9 in Eastern Europe, $4–7 in Latin America and the Caribbean, $2–3 in the Philippines and $1.50–2.50 in India.

The gap between those wages and your invoice is the provider's facilities, management, technology, quality function and margin — and it varies enormously between a 446,000-person incumbent and a 4,000-person provider. Every "$12–18 nearshore" figure in circulation traces back to provider marketing. Negotiate against your own cost per resolved contact instead, and ask any provider quoting a rate to name the delivery location it assumes.

Is the customer service outsourcing market growing?

The forecasts and the filings disagree. Grand View Research values customer experience BPO at $113.0 billion in 2025 rising to $296.3 billion by 2033, a compound rate of 13.0%. Against that, ISG recorded total business process outsourcing annual contract value down 14% in FY2025 to its lowest since 2020.

The audited numbers from the providers themselves are more pointed. Concentrix impaired $2.57 billion of goodwill across four quarters, the second tranche explicitly attributed to its own share price. Teleperformance has posted four consecutive declining periods. TELUS booked a $500 million impairment against its digital experience unit months before taking it private. The work is not disappearing, but the dollars attached to each unit of it clearly are.

Should I choose a large incumbent or a mid-sized provider?

Work out whether you are buying scale or attention, because the answer eliminates most of this list immediately. If you need 5,000 seats across twenty languages with mature global procurement, the mid-sized providers cannot serve you at any price and Concentrix or Teleperformance are the answer. If you need 200 seats and real operational attention, the incumbents structurally cannot give it to you.

The arithmetic is simple enough to state: a 200-seat programme is roughly 0.04% of a 446,716-person provider's headcount. It will get a shared account manager and a standard playbook, and that is a rational response to its size rather than a failing. The same programme is about 5% of a 4,000-person provider and will get named leadership because it has to.

What happened to Foundever in 2026?

It closed a recapitalisation in August 2026 that cut nearly $900 million of debt and brought in $225 million of common equity from its existing majority shareholders, with 95.4% of term loan lenders and 100% of revolving credit lenders participating. Both Laurent Uberti, the group chief executive, and Olivier Camino, the executive chairman and deputy chief executive, stepped down with effect from 31 July 2026, with an interim chief executive in post.

Headcount has contracted roughly 24% from about 170,000 at the 2023 rebrand to about 130,000, and its last public revenue figure remains $4 billion as of March 2023. Reporting on the recapitalisation cited clients building support in-house with AI instead of outsourcing as a cause of the distress. It remains an Everest Leader for both 2025 and 2026, which is the clearest illustration on this page that analyst standing and financial health are different things.

Which providers have changed ownership or listing status recently?

Four of the twelve, and most competing rankings have not caught up. TELUS completed the privatisation of TELUS Digital on 31 October 2025 at roughly US$539 million, delisting it from the NYSE and TSX. Startek, taken private by Capital Square Partners in 2023, completed a merger with CCI Global on 16 April 2026 and disappeared from Everest's 2026 matrix as a result.

Foundever recapitalised in August 2026 as described above. HGS divested its healthcare BPO arm — 20,000-plus staff and roughly $400 million of revenue — to Baring Private Equity Asia at a $1.2 billion enterprise value in 2021, and that business now trades as Sagility. What remains of HGS is a much smaller company than its brand history suggests.

What is the difference between customer service outsourcing and BPO?

Business process outsourcing is the broad category: any business process run by an external provider, including finance and accounting, human resources, procurement, claims processing and IT support. Customer service outsourcing is the subset that handles interactions with a company's own customers — voice, chat, email, social and the back office work behind a resolved contact.

The distinction matters when reading market data, because the two are frequently conflated. ISG's annual contract value figures cover all BPO, so a 14% decline there is not a customer-service-specific signal. Grand View's customer experience BPO sizing is narrower. When a provider quotes you a market growth figure, ask which of the two it is measuring.

Will AI replace outsourced customer service?

Not on the measured evidence, but it is already restructuring who supplies it. Gartner predicts agentic AI will autonomously resolve 80% of common service issues by 2029 and Salesforce expects half of all cases AI-resolved by 2027. ContactBabel's measured figure is 18% of web chats handled with no human agent in 2026, up from 6% in 2020, with about 22% still escalating to another channel.

Consumer appetite runs the other way: Gartner found 64% of 5,728 consumers would prefer companies did not use AI for customer service and 53% would consider switching over it. What is unambiguous is the supplier-side effect — ISG attributes falling contract value to AI compressing what each deal is worth, and Foundever's lenders cited client insourcing with AI as a cause of its distress. Ask every provider what share of its revenue sits with clients who have announced AI-first strategies.

What certifications should a customer service outsourcing provider have?

SOC 2 Type II and ISO 27001 in its current 2022 revision are the baseline, with PCI DSS where cardholder data is in scope and ISO 27701 where privacy is. ibex is the only provider on this page certified to ISO/IEC 42001, the AI management system standard, which is becoming relevant as AI enters delivery.

Five words matter and are not interchangeable: certified means an independent auditor assessed the organisation and a certificate exists; compliant means self-attested against a standard with no certifying body; aligned means controls mapped but not attested; claimed means asserted without independent confirmation; not held means absent from every source reviewed. There is no such thing as HIPAA certification. With PCI DSS, ask which of the four levels — only Level 1 requires a QSA assessment and an Attestation of Compliance — and then ask about scope.

Which provider is growing fastest?

ibex, on disclosed numbers: FY2026 revenue of $644.1 million, up 15.4%, with net income of $46.3 million. It is the only listed provider on this page growing double digits and profitable, in a year when Concentrix impaired $2.57 billion and Teleperformance posted four consecutive declining periods.

That comes with a caveat worth checking. Record growth at roughly 35,000 staff usually means a small number of large accounts, so ask for the client concentration disclosure in the annual filing before treating the growth as a stability signal. Everest rates ibex a Major Contender rather than a Leader, and it cannot serve the very largest consolidations.

How long does it take to launch an outsourced customer service programme?

There is no independent benchmark and every ramp figure in this category is a vendor claim, so this page reports them as claims. The spread on this page is wide: the large incumbents score between 4.8 and 6.2 on speed to launch because they are structured around multi-year enterprise agreements, while Helpware publishes a 30 to 60 day proof of concept with a path to 500-plus FTE in 90 to 120 days.

The difference reflects what each provider is built for rather than how competent it is. A twenty-country consolidation should take months and anyone promising otherwise is overselling. A 50-seat pilot should not. Write whichever ramp you are promised into the statement of work with milestone dates and named people.

What should I ask about a provider's financial health?

In 2026 this belongs in your selection criteria rather than your footnotes, and the question is harder than it looks because six of the twelve providers on this page disclose nothing at all — no revenue, no net income, no ownership percentage. Sutherland and Transcom do not even publish a headquarters city.

For the listed ones, read the filing rather than the press release. Concentrix's $2.57 billion of impairments across four quarters, Teleperformance's four consecutive declining periods and HGS's standalone loss of ₹160.58 crore are all in the public record. For the private ones, ask for audited revenue and net income under NDA and treat a refusal as information. Also ask what share of revenue sits with clients announcing AI-first support strategies, which is the exposure that cost Foundever its leadership.

Which providers are best for European delivery?

Transcom holds Everest EMEA Leader and Star Performer standing — a better placement than several larger providers — with 29 countries and 80-plus delivery centres at a cost base the giants cannot match. Foundever has the deeper footprint at 45-plus countries and 60-plus languages and an Everest Leader placement, with the financial caveats described above.

For Eastern European delivery specifically, Helpware blends it with onshore US and Philippine capacity inside one contract, which is the pattern that suits a US buyer wanting timezone overlap rather than pure offshore cost. The published wage band for Eastern Europe is $5–9 an hour against $17–22 onshore, but remember those are wages rather than the rate you will be billed.

How was this ranking produced, and who paid for it?

Nobody paid for it. No provider was invited to participate, shown its score in advance, or given a chance to respond before publication, and no provider can pay to be included, ranked higher, described differently or removed. There is no advertising, no affiliate compensation and no referral arrangement with any ranked provider.

The method is published above: eight weighted criteria disclosed before any result, composites computed in code rather than assigned by hand, and the full per-criterion matrix printed so that any reader can disagree with an individual judgment and recompute the order. Every external figure links to its primary source at the point of use, and no statistic anywhere on this page is sourced to a provider's own marketing material — including the provider that ranks first, whose own Clutch claim this page contradicts using the live figure.

Sourcing

Where the numbers come from

Every statistic on this page links to its primary source at the point it is used. Those sources are SEC filings read directly from EDGAR, Euronext and Indian exchange filings, Everest Group's CXM Services PEAK Matrix, the ISG Index, Grand View Research, Site Selection Group, ContactBabel's US Contact Center Decision-Makers' Guide, SQM Group, Gartner, Salesforce, the US Bureau of Labor Statistics, and company announcements for the corporate events described. Review ratings come from Clutch and G2 profiles read live on 7 October 2026. There is no aggregated bibliography at the foot of this page because a duplicate list adds length without adding verifiability — follow the link where you meet the number.

No statistic on this page is sourced to a customer service provider's own marketing material, including the provider that ranks first — whose own Clutch claim this page contradicts using the live figure. Vendor-reported outcomes are labelled as such and unaudited in the same block they appear in. Where no independent source exists for something a reader would reasonably want, the evidence section names the gap rather than substituting a weaker number: there are no independent billable rates by region, no outsourced-versus-in-house quality comparison, and no ramp-time benchmark.