Displacement or Disruption? Manila’s Outsourcing Sector Faces Its AI Reckoning

Lisa trained the AI system that replaced her, then watched her job disappear one month before it was due to become permanent, according to an account she gave the BBC on condition of anonymity.

Her story sits inside a wider argument the Philippines has not settled: the same category of routine, entry-level digital work that labor researchers warn is disappearing is also the exact work a booming virtual-assistant industry claims to be moving past.

Two Ledgers, One Labor Market

Two accountants can look at the same ledger and file two different reports. One sees a business bleeding out, quarter over quarter. The other sees a portfolio posting its best numbers yet. Both are reading real entries. Neither needs to be lying. They’ve simply been handed different columns.

That is roughly the shape of the argument now underway over what generative AI is doing to Philippine outsourcing work.

On one side, the International Labour Organization estimates that 12.7 million Filipino workers, more than one in four, hold jobs with some exposure to generative AI, the highest share in Southeast Asia, and workers like Lisa describe being made redundant after their writing was used to train the tools that replaced them.

On the other, a recruitment agency’s data, republished across at least five outlets this month, describes a virtual-assistant sector so oversubscribed it places fewer than 1 in 500 applicants, and frames its growth as proof that AI concerns are overstated.

Neither account cites the other’s numbers. Both are describing the same shrinking category: routine, admin-heavy digital work, the kind that gets automated first and marketed around second.

The Stories behind the Numbers

12.7 million is a plain figure. It gets filed the way population counts and rainfall totals get filed, a line in a report, sourced to the ILO’s 2025 global assessment. It does not record severance pay. It does not record the eight months Lisa spent at her employer, or the one month left before her contract converted, or the fact that she says a public relations agency was brought in specifically to generate the material she and her colleagues were then asked to edit.

Mary, a second former content writer who spoke to the BBC under a changed name, describes a similar arc: AI tools that were meant to save time instead required more fact-checking, more editing, and, eventually, her own redundancy.

Industry figures push back on the direct-causation reading.

Jack Madrid, president of the IT and Business Process Association of the Philippines, told the BBC that more than two-thirds of the group’s members are already running AI pilots, but that most displaced workers have been redeployed rather than let go, and that any hiring slowdown reflects weaker global demand as much as automation.

Paul Quintos, of the University of the Philippines-Diliman, goes further, telling the BBC that “AI washing” is now a recognizable pattern: layoffs driven by ordinary demand softness get relabeled as AI-driven progress, because that framing reads better than an admission that the market slowed down.

Which explanation fits any given layoff, Lisa’s included, isn’t resolved.

The Industry That Says the Opposite

Set against that is a very different account of the same sector, and it comes from a single source.

Five outlets this August 2026, including Manila Bulletin, Sun Star Davao, Outsource Accelerator, and Allwork.Space, published figures from the same proprietary report by VA Masters, a global recruitment agency, which puts the global virtual-assistant market at $6.5 billion in 2026, headed toward $34 billion by 2030, with the Philippines supplying an estimated 38% of that workforce.

VA Masters CEO Alon Pearl, in comments distributed to press, framed the pitch around selectivity rather than cost: “Everyone talks about the Philippines as a source of affordable talent,” he said. “What the data actually shows is a talent pool so deep that we place fewer than 1 in 500 applicants. The story here isn’t cost, it’s selectivity.”

The figures are consistent on the traits that matter to that pitch:

• 69% of placed VAs are women, against 61.5% of applicants
• 80% of hired workers report fluent or native-level English, against 62% of the applicant pool
• The median placed worker is 28

We can’t help but note, however, that the underlying applicant count itself is not consistent across publications. Two of the five outlets cite 166,350 applications; the company’s own report and two others cite 102,677 for what is described as the same 12-month period. Seems rather specific for a typo, and due to the nature of the disparity, quite attention-grabbing.

And nowhere in that coverage does the growth narrative engage the ILO or BBC reporting directly, beyond a single passing line in one Manila Bulletin piece acknowledging AI concerns before pivoting back to signup numbers.

What’s Unreconciled from Neither Side

Read the two accounts side by side and they are not actually arguing about which jobs are exposed.

A Sun Star report on the VA industry’s own growth says demand is shifting “well beyond basic administrative tasks,” toward specialized customer service, technical support and platform management, and explicitly names data entry, clerical work and cold calling as the tasks “more vulnerable to automation.”

That is, functionally, the same category of work Lisa and Mary describe losing.

The industry says it is moving past that work. The workers who did it say it is what they were made redundant from. Both statements can be true, and if they are, the honest headline is not “AI is destroying Philippine jobs” or “AI concerns are overblown.” It is that the bottom rung of the outsourcing ladder is being retired by two industries with no tacit agreement, nor notification to the people standing on it, where it goes.

The Philippines’ Department of Information and Communications Technology has committed to upskilling more than 300,000 outsourcing workers and has explicitly declined to push new AI-specific legislation, arguing existing labor and privacy law can be adapted faster through guidance.

Whether that reaches the Lisas and Marys of the sector, rather than the workers already positioned to move into the specialized roles the VA industry is expanding, is the open question that no coverage mentioned here — ILO’s or VA Masters’ — is currently built to answer.