Confidential Strategy Briefing · TP Malaysia & Singapore

TP's €100M AI bet gets won on the floor — in Penang and KL.

FY2025: €10.2B revenue, 14.6% margin, and a new ex-McKinsey CEO betting the company on AI-native operations. This briefing maps where that bet quietly leaks value at site level, and how to recover it without the morale and CSAT hit that usually follows AI-led cost cuts.

Margin inflection — TP site economics
LEGACY SEAT-BASED REVENUEAI-DEFLECTED VOLUMEYOU ARE HEREOUTCOME + HUMAN-PREMIUMVALUE (recoverable)
As AI absorbs volume and legacy seats fall, margin either collapses or climbs the green path. The green path has no owner yet.
GROUP REV
€10.2B
EBITA MARGIN
14.6%
AI SAVINGS TARGET
€100M+
SHARE / 12 MO
−49%
01Current State

Where TP stands today

A market leader — and a company under real pressure to convert AI ambition into margin.

Group · Teleperformance SE (FY2025)

REVENUE FY25
€10.2B
RECURRING EBITA MARGIN
14.6%
NET FREE CASH FLOW
€901M
EMPLOYEES
~447K
COUNTRIES
~100
LANGUAGES
300+
FINANCIAL SVCS 41%GOV / ADMIN 40%TELECOM 12%RETAIL 7%
14.6% EBITA margin is solid, but competitive benchmarks in AI-native BPO (Atos, Wipro onshore) are pushing 16–18%. TP has to move fast without breaking the talent machine in Malaysia—where churn is tied to perceived job security.

TP in Malaysia (current)

ESTABLISHED
2017
SITES
5 + JB
EMPLOYEES
2,200+
LANGUAGES & DIALECTS
25+
GLOBAL MARKETS SERVED
15+
F&S MARKET LEADERSHIP
'25

3 Penang + 2 KL sites, Johor Bahru incoming · multiple 2025 CCAM Awards incl. Best Use of Automation · only BPO first-mover in Penang.

02Where Value Leaks

Three areas mapped in research

Each is a mechanism, not a complaint — and each has a recoverable upside with a specific lever.

03The Next AI Move

Where Jorge Amar is taking TP

The Future Forward plan re-points the whole company at AI-native operations: the TP.ai platform with its agent-building "Fab," agentic-AI partnerships with Ema and Parloa, and real-time speech/accent via Sanas. A Chief AI Officer role is being scaled, and the billing model is shifting from headcount / seat toward outcome / value-based pricing — alongside new AI data-services revenue and a 2028 target of ~15.5% EBITA margin post-transformation.

Future ForwardTP.ai platformAgent-building "Fab"EmaParloa (agentic AI)Sanas (real-time speech)Chief AI OfficerSeat → outcome billingAI data services2028 target ~15.5% EBITA

“The whole company is being re-pointed at AI-native operations. The mandate is set. The question is who delivers it on each floor.”

04The Gap

Where the ambition meets reality

  • Fragmented AI integration → broken customer journeys (client-reported).

  • 30–40% attrition draining margin and SLA stability.

  • Monitoring/surveillance backlash on the floor — every action tracked, restrictive breaks → morale and retention cost. (Public employee reviews.)

  • Workforce not upskilled fast enough for tier-2/3 complexity → burnout.

  • Trust & Safety psychological toll → wellness / duty-of-care exposure.

  • Language-premium friction between local and expatriate native speakers.

The structural risk

If AI reduces billable volume faster than TP shifts to outcome-based contracts, revenue-per-employee falls and margin recovery lags. This is the whole game.

None of this is a tooling problem. It's an ownership problem — the seam between Ops, Tech and People has no owner.

05What I Offer

The work nobody currently owns

The COO owns SLAs, the CIO owns the stack, HR owns headcount — but the seam where AI + operations + workforce transition meet is exactly where margin leaks, and it has no owner.

Proposed role: Lead — CX-AI Transformation & Workforce Transition (Malaysia & Singapore).

I'm proposing to prove the role by doing it: start as a paid diagnostic, convert to the mandate.

I select, integrate, and drive adoption of proven tools and wire them into your CRMs and floor. I don't build models.

How we start

  1. 01Weeks 0–3

    Diagnostic

    Quantify the value pools on your data; pick one lighthouse. Low cost, high trust.

  2. 02Days 30–90

    Lighthouse win

    One site, one use case (likely predictive attrition or copilot). Prove ROI in hard numbers.

  3. 03Q2+

    Scale + own the transition

    Expand the win; own the seam between Ops, Tech and People.