The Philippine IT-BPM industry cut its own 2028 forecast in July 2026, and it named AI as a lead reason. IBPAP now expects industry revenue of $43.3 billion to $50.5 billion by 2028, down from the $59 billion target set in 2022, with the employment goal trimmed from 2.5 million to between 1.85 million and 2.14 million. A revision that large invites a practical question that most coverage skips. Which of the country's BPO companies are actually running AI in production, and which are still announcing pilots?
This benchmark gathers the production claims that named operators in the Philippines have published, each with a date, a source, and a confidence label. One rule applies throughout: figures from company press releases and investor decks are self-attested, not verified, and they are marked that way in the text. That distinction is the current state of evidence on AI in BPO, and pretending otherwise helps nobody.
The industry repriced itself before displacement showed up in the data
The macro numbers set the stage. The sector closed 2025 with 1.9 million workers and $40 billion in export revenue, equivalent to more than 8 percent of GDP. For 2026, IBPAP projects revenue of $42.3 billion and employment of 1.96 million. Growth continues, in other words, just more slowly than the 2022 roadmap assumed.
What has not yet appeared is mass displacement. IBPAP president Jack Madrid put it plainly when the revised targets were announced: "I think AI is a real development, but I think we have not really seen it scale yet." The association is pairing the downgrade with skills spending, including Project UNLAD, a P740 million training partnership with DICT and TESDA. The gap between a repriced forecast and limited observed displacement is exactly why company-level production evidence matters.
What the major providers say they are running
Concentrix describes iX Hero as an agentic AI platform embedded directly into advisor workflows. The company reports that over 12 months the platform cut average handle time by 22 percent, reduced advisor search time by 20 percent, and lifted customer satisfaction by 13.5 percent. These figures are self-attested as of 6 August 2026 and carry no Philippines-specific breakdown.
iQor, which runs approximately 45,000 employees across 10 countries, said in June 2026 that its infinityAiQ platform lets clients shift up to 30 percent of transactional volume to agentic solutions, and that over 60 percent of its business is now enabled by AI. Both claims are self-attested as of 10 June 2026.
Alorica reported in its 2025 year-end review that the majority of new logo opportunities in 2025 included an Alorica IQ component, the AI and analytics stack that includes its evoAI conversational platform. The company counts more than 100,000 staff across 16 countries and names the Philippines among its growth markets, but publishes no country-level AI metrics. Self-attested as of 18 March 2026.
TELUS Digital reports that its Fuel iX platform processed more than two trillion tokens in 2025 across more than 20 generative AI use cases scaled into production, including simulated-voice agent training and live agent assist. These are company-reported figures, self-attested as of 25 February 2026.
Teleperformance, the largest of the group, told investors its TP.ai FAB suite generated 550+ AI projects and 50+ won deals in the first quarter of 2026 alone. Self-attested as of 28 April 2026.
Audited filings put harder numbers behind two of the claims
TaskUs is the clearest case where AI shows up in the accounts rather than the marketing. The company closed 2025 with record revenue of $1.18 billion, up 19 percent year over year, with AI services growing nearly 59 percent for the full year and 46 percent in the fourth quarter. It plans to invest over $25 million in AI transformation and emerging growth initiatives in 2026. Earnings figures sit a step above press-release claims because they pass through audit and disclosure rules.
Teleperformance's own results show why deal counts alone do not settle the question. The same first-quarter presentation that reported 550+ AI projects also reported group revenue of 2.433 billion euros, down 2.2 percent like-for-like. AI momentum and soft core revenue are coexisting inside the BPO industry's biggest players, a pattern consistent with clients redirecting spend rather than simply adding it.
Philippine capacity keeps growing while the AI reviews run
Production AI has not stopped physical expansion. TP Philippines, with over 47,000 employees, will open its 29th site in Pasay City in September 2026, adding more than 1,000 jobs under a PEZA supplemental agreement signed on 29 June 2026.
The caution is visible too. In November 2025, Teleperformance deputy CEO Thomas Mackenbrock, asked whether AI would cut jobs after Accenture announced mass layoffs that September, said the group was looking "step by step, process by process, structure by structure, at how we can be better set up for the future and drive this with AI adoption." Hiring and restructuring at the same time is what this transition looks like from the inside.
How to read self-attested claims, and what to ask for next
Every operational statistic in the provider section above comes from the company itself. None has been independently audited. Our scoring reflects that: registry and official records score 0.90 to 0.98, third-party reporting 0.75 to 0.89, and self-attested company claims 0.50 to 0.74. A press release that names a tool, a use case, and a dated metric is evidence, but it stays weak evidence until a client, an auditor, or a regulatory filing corroborates it.
For buyers, policymakers, and researchers, the working test is short. Ask for four things: the named tool, the named process it runs inside, the date it went live, and the metric with its baseline. Providers with genuine deployments can answer in a sentence. We maintain company-level AI-in-production flags built on this standard across the 943 profiles in our BPO directory, each carrying the same confidence scoring used in this article.
