Teleperformance

Teleperformance scores 78.8 out of 100 and sits in the Leader tier of both editions. The score comes from a published 100-point framework applied to public sources rather than a client engagement, and it rests on 13 evidence claims, each graded and sourced below.

In its own words“TP is a global partner that operates where people, processes, and AI meet.”

Key facts

Headquarters
France: Paris
Operating since
1978
Workforce
~490,000 employees at 31 Dec 2025
Revenue
2023 revenue €8,345M, net profit €602M
Languages claimed
400+ languages and dialects group-wide on TP’s own About Us page; its 2024 Integrated Report states around 300. No European-only language count is published
Delivery footprint
Pan-European + global, 90+ countries
Employee signal
4.4 across 75,000+ reviews, pulled 22 Aug 2026 up 2% over twelve months; compensation rating 4.3; management-response activity present
Commercial entry
Enterprise contract
AI capability
TP.ai Connect · TP.ai Assist · TP.ai Collect · TP.ai Growth · TP.ai Dataservices
What carried the score

Scale + capability sweep

What held it back

Certification scope per entity is not publicly traceable; client-side review evidence does not exist at this scale

Evidence was last accessed on 17 August 2026. Scores are final for the 2026 Edition, and every figure here can be checked against the evidence further down this page.

When to shortlist

One published mapping, applied to every provider the same way. A score of 4 or 5 on a situation’s sub-criterion becomes something you can rely on; 2 or below is something you would have to establish yourself. Verified languages come straight from the data sheet. At most two of each appear, in a fixed order. See all situations.

  • You can rely onClaims independently checkable A5.1 = 4Material claims are checkable and consistent across two or more independent sources.
  • You can rely onAI in production B6.2 = 4Runs AI in production with the workflows and the tools named.
  • Check for yourselfVerified language delivery B2.1Fewer than five languages, or fewer than three non-English European ones, are verifiable against hiring or client evidence.
  • Check for yourselfNamed client cases A1.3 = 2Publishes no case study that names its client.

Company background

Every fact below is drawn from the evidence listed further down, with its source and grade, in the order it appears here.

Identity & origin

Paris head office, founded 1978; listed on Euronext Paris in the CAC 40.

Scale & footprint

It states approximately 490,000 employees at 31 December 2025. The most recent revenue figure held is 2023: 8,345M, with net profit of 602M.

Scale figures vary across Teleperformance's own materials and the variance is the finding: the 2025 Integrated Report states nearly 490,000 and about 170 markets, while the About Us page markets 400+ languages and dialects against around 300 in the 2024 Integrated Report.

What the provider says

Its published scope covers trust and safety, collections, healthcare support, visa and consular services and interpretation, across 13 named industry verticals. European delivery spans Poland, Portugal, Greece, Germany, France, Spain, Italy, the Netherlands, Turkey, the UK and the Nordics, with Greece stated at more than 8,700 employees.

Third-party evaluation

Everest Group names it a Global Leader in the CXM PEAK Matrix, recurring. Great Place To Work certification rests on two dated company announcements rather than one current figure: 72 countries as at July 2023 and 69 as at July 2024, covering 97% of employees, with the 2024 survey data collected before the Majorel acquisition completed. Glassdoor reads 4.4 across 75,000+ reviews, pulled 22 August 2026.

Not evidenced

A dedicated security section exists on tp.com; per-entity certificate numbers are not published there and are not traceable to a registry from public materials.

Commercial entry

Enterprise contract.

How Teleperformance's stated scale compares across the field: the 29-provider breakdown by evidence quality. The five closest-scoring alternatives to Teleperformance: on the same twelve criteria.

Strengths and weaknesses

These are the sub-criteria where Teleperformance sits furthest above and below the median of the 29 providers in the Global Edition. Selection is by distance from the field, not by raw score, because a 5 where most of the field scores 5 says nothing.

These are scores for what a provider publishes, so a low score can mean a weak record or a quiet one, and this comparison can’t tell them apart.

Furthest above the field

  • Employee-side signal A4.15 vs field 1
  • Payment & data-handling compliance A3.24 vs field 2
  • Native-language delivery, verified B2.15 vs field 3

Furthest below the field

No sub-criterion sits below the middle of the field.

How we scored it

Each sub-criterion gets an anchor from 0 to 5 against a published description. The anchor divided by 5 and multiplied by the sub-criterion’s points is its contribution, and the contributions add up to the total. Scored against version v1.2-c of the published framework.

CodeCriterion and sub-criterion anchorsShare of points earnedPoints
A1Independent client evidenceA1.1: 3, A1.2: 3, A1.3: 2Analyst-equivalence rule applied (reference checks within Everest assessment); cases largely anonymized.7.6 / 14
A2Market presence & tenureA2.1: 5, A2.2: 5, A2.3: 58 / 8
A3Certifications & complianceA3.1: 3, A3.2: 4, A3.4: 3Restructured A3. Scope-per-entity unverified.8 / 12
A4Delivery stabilityA4.1: 5, A4.2: 5Glassdoor profile E9779 (glassdoor.com/Reviews/Teleperformance-Reviews-E9779.htm): 4.4 across 75,000+ reviews, above the anchor's 4.0 threshold across at least 25 reviews, with management-response activity present. The anchor's client-review cross-check is not applicable: this provider has no client-review footprint that could contradict it.6 / 6
A5Transparency & verifiabilityA5.1: 4, A5.2: 3Filings strong; headcount variance noted; pricing opaque.3.6 / 5
B1Service scope & channelsB1.1: 5, B1.2: 5, B1.3: 58 / 8
B2Language & geographic reachB2.1: 5, B2.2: 5, B2.3: 510 / 10
B3Quality management systemB3.1: 4, B3.2: 3, B3.3: 47.4 / 10
B4Security & data protection operationsB4.1: 4, B4.2: 3, B4.3: 56.6 / 8
B5Flexibility & commercial modelB5.1: 1, B5.2: 4, B5.3: 12.6 / 7
B6Technology & AI operationsB6.1: 5, B6.2: 4, B6.3: 4B6.2 held at 4 not 5: AI outcome numbers are mostly TP-stated, not client-attributed.6 / 7
B7Vertical expertiseB7.1: 5, B7.2: 55 / 5

Delivery Proof adds up to 33.2 of 45 and Capability Fit to 45.6 of 55, for a total of 78.8 of 100. The rule is printed in the scoring anchors appendix.

AI capabilities

ProductAs the provider describes it
TP.ai ConnectHybrid intelligence orchestrating people, AI, and workflows
TP.ai AssistAI-powered assistant driving operational scale
TP.ai CollectAI-powered collections for faster, smarter recovery
TP.ai GrowthAI-powered sales operations that accelerate B2B revenue
TP.ai DataservicesNo description is published.

A “Responsible AI control center” is described as providing “full visibility, traceability, and auditability” with “built-in governance and compliance by design”, and the orchestration engine “keeps humans in control where precision matters”.

No sampling rate, escalation trigger or containment-versus-deflection definition is published, so the governance is stated at the level of a capability rather than a design. Checked 31 Aug 2026.

how many of the 32 providers assessed publish a control over their AI

This evidence was gathered in a supplementary pass on 30–31 August 2026 and is not reflected in any score.

Evidence

Each claim carries a grade, and the grade decides what a score can reach. A sub-criterion scored 5 needs at least one E1 or E2 source; evidence that is entirely company-stated caps it at 3. This sheet holds 13 claims, and a claim citing more than one grade is counted under each, so the counts exceed 13.

E1 3 E2 2 E3 7 E5 2

  1. #0E1

    Paris HQ, founded 1978; Euronext Paris, CAC 40

    Source: Euronext listing page for Teleperformance SE (FR0000051807, XPAR); read 28 Aug 2026. Confirms the Paris listing and CAC 40 membership. Founding year 1978 appears throughout the 2024 Universal Registration Document Open the source

  2. #1E3

    Scale figures vary across TP’s own materials and the variance is the finding: the 2025 Integrated Report states “nearly 490,000” under Human capital and about 170 markets; the About Us page markets 400+ languages and dialects while the 2024 Integrated Report states around 300. Post-Majorel (November 2023) figures have not settled to one published set

    Source: Teleperformance Integrated Report 2025 and tp.com About Us, checked 28 Aug 2026. Company-stated throughout, so E3: every figure here is TP’s own and the conflict is between its own documents. The report renders its headcount as an infographic whose unit label is not machine-readable, so the noun attached to 490,000 is not quoted Open the source

  3. #2E1

    2023 revenue €8,345M, net profit €602M

    Source: Teleperformance SE Universal Registration Document 2024, which carries the 2023 comparatives; read 28 Aug 2026. The €8,345M revenue figure appears in the consolidated accounts Open the source

  4. #3E1

    Everest CXM PEAK Matrix Global Leader, recurring

    Source: TP press release on the Everest Group CXM Services PEAK Matrix Global 2025; read 28 Aug 2026. It states TP is the highest-designated Leader in Market Impact and a Star Performer, and a Leader across Americas, EMEA and APAC Open the source

  5. #4E3

    Full CX scope including trust and safety, collections, healthcare support, visa and consular services and interpretation. The TP.ai FAB suite spans Assist, Connect, Growth, Data, Collect and Operate, with 550+ AI projects reported in Q1 2026

    Source: Teleperformance FY2025 and Q1 2026 investor presentations, checked 28 Aug 2026. Company-stated: the suite composition and the project count are TP’s own, and the outcome figures attached to individual FAB products are company-reported with no named client Open the source

  6. #5E3

    13 named industry verticals: automotive, banking and financial services, crypto, energy and utilities, healthcare, insurance, media, public sector, retail and e-commerce, technology, telecom, travel/hospitality/cargo, and video games

    Source: tp.com industry pages, checked 28 Aug 2026. Company-stated. This claim previously said 12; TP’s own pages name 13 Open the source

  7. #6E5

    Dedicated “Security and process excellence” section on tp.com; per-entity certificate numbers are not published there or traceable to a registry from public materials

    Source: tp.com “Security and process excellence” section, and a search for the named standards in public certification-body client directories; no per-entity certificate number or scope located. checked during our August 2026 evidence window

  8. #7E2

    Great Place To Work certification, on two dated company announcements rather than one current figure: 72 countries as at July 2023, and 69 countries as at July 2024 covering 97% of employees. The 2024 survey data was collected before the Majorel acquisition completed on 8 November 2023, so neither figure reflects the combined workforce

    Source: Teleperformance announcements via businesswire.com and prnewswire.com, 5 Jul 2023 and 16 Jul 2024, checked 28 Aug 2026. E2: Great Place To Work is a third-party certifier and the certification is its determination rather than the provider’s. The country counts are not a range within one year; they are two separate annual results, and the claim previously presented them as “69–72” Open the source

  9. #8E2

    Glassdoor 4.4 across 75,000+ reviews, pulled 22 Aug 2026; up 2% over twelve months; compensation rating 4.3; management-response activity present

    Source: Glassdoor employer profile E9779 (TP/Teleperformance); recorded 22 Aug 2026. Figures as recorded in the collection window; not re-checked since. Entity confirmed 28 Aug 2026 via the profile listing, which carries the same 4.4 rating and 4.3 compensation figure; the review count has since risen past the 75,000+ recorded here Open the source

  10. #9E3E5

    European delivery across country operations including Poland, Portugal, Greece, Germany, France, Spain, Italy, the Netherlands, Turkey, the UK and the Nordics. Greece is stated at more than 8,700 employees across Athens, Thessaloniki and Chania; Portugal has grown past 10,000. Operations in Romania, Ukraine, Croatia, the Czech Republic and Lithuania were previously listed here but are not confirmed in TP’s current published country coverage

    Source: Split grade. E3 for the confirmed country operations, from TP’s own country and careers pages and its European workplace award coverage, checked 28 Aug 2026. E5 for the five countries removed; a buyer needing delivery in any of them should confirm it directly rather than infer it from a group footprint Open the source

  11. #10E3

    TP.ai Connect, Assist, Collect, Growth and Dataservices are named on the AI page, which publishes 80% QA reduction, roughly 60% faster deployments and roughly 70% frictionless resolution, with one named case: Sparda-Bank West at 46% first-contact resolution. A “Responsible AI control center” is described as providing “full visibility, traceability, and auditability” with “built-in governance and compliance by design”, and the orchestration engine “keeps humans in control where precision matters”. No sampling rate, escalation trigger or containment-versus-deflection definition is published, so the governance is stated at the level of a capability rather than a design. Checked 31 Aug 2026.

    Source: tp.com/ai, checked 31 Aug 2026 Open the source

  12. #11E3

    Each named TP.ai product carries a one-line description on the AI page, quoted here verbatim: TP.ai Connect: “Hybrid intelligence orchestrating people, AI, and workflows”; TP.ai Assist: “AI-powered assistant driving operational scale.”; TP.ai Collect: “AI-powered collections for faster, smarter recovery.”; TP.ai Growth: “AI-powered sales operations that accelerate B2B revenue.” TP.ai Dataservices is named on the same page with no descriptive line, so none is recorded for it (an absence in the source, not an omission here). These are the company’s own product taglines rather than functional specifications, and no scope, model, deployment detail or evaluation is published alongside them. Checked 31 Aug 2026.

    Source: tp.com/ai product descriptions, checked 31 Aug 2026 Open the source

  13. #12E3

    POSITIONING, as the provider states it: “TP is a global partner that operates where people, processes, and AI meet.”. Company-stated, so E3: this records what the provider says about itself, not what the evidence establishes.

    Source: www.tp.com homepage, checked 31 Aug 2026 Open the source

Open questions

The research could not settle these from public sources. Put them to the provider directly, alongside the 18 questions for every shortlist.

Not published, so verify directly

  • Dedicated “Security and process excellence” section on tp.com; per-entity certificate numbers are not published there or traceable to a registry from public materials #6
  • European delivery across country operations including Poland, Portugal, Greece, Germany, France, Spain, Italy, the Netherlands, Turkey, the UK and the Nordics. Greece is stated at more than 8,700 employees across Athens, Thessaloniki and Chania; Portugal has grown past 10,000. Operations in Romania, Ukraine, Croatia, the Czech Republic and Lithuania were previously listed here but are not confirmed in TP’s current published country coverage #9

Flags and unresolved readings

  • Certification scope per entity is not publicly traceable (E5); a buyer should request it directly.
  • A4.1 is now anchored on a recorded rating (Glassdoor 4.4 across 75,000+ reviews, 22 Aug 2026) rather than on an expectation.
  • B6.2 held at 4 not 5 because AI outcome numbers are mostly TP-stated, not client-attributed.

Certifications and verticals

Verticals claimed

  • trust and safety
  • healthcare
  • collections

These lists are as the provider names them. They are recorded for reference and are not scored on their own.

Report an error

Every line on this page is meant to be checkable. If a datapoint is wrong, out of date, or missing a certificate identifier we could not find, the corrections process is open. Submit the evidence and the score is re-derived through the anchors. Facts can be corrected, and scores follow the facts.