OLKERIAI News
← All AI news
The Philippines Runs the World's Call Centres. AI Is Coming for Them

Image: Olkeri

Society & CultureAsia29 August 20263 min read

By Olkeri.space

The Philippines Runs the World's Call Centres. AI Is Coming for Them

An economy built on voice services faces the most direct AI disruption of any country, and is racing to move up the value chain.

Read this story in: Français · Deutsch · Español

The Philippines faces a more direct confrontation with artificial intelligence than almost any other economy, because the industry AI most readily automates is one of its largest employers.

The outsourcing economy:

Business process outsourcing employs well over a million Filipinos and generates a very large share of export revenue. The country became the world's leading destination for voice-based customer service, built on widespread English proficiency, cultural familiarity with American norms, and a large young workforce.

Beyond call centres, the sector includes back-office processing, healthcare administration, finance and accounting, and increasingly higher-value analytical work.

The direct threat:

Conversational AI systems handle a growing share of routine customer interactions competently. The tasks most susceptible, answering standard questions, processing simple requests, following scripted resolution paths, are precisely what entry-level outsourcing roles involve.

Industry response has been to reposition rather than resist: emphasising complex, emotionally sensitive and high-value interactions that AI handles poorly, and moving into AI-adjacent work including data annotation, model evaluation, quality assurance for AI outputs, and managing AI systems.

That transition is genuine and incomplete. The higher-value work requires different skills and employs fewer people per unit of revenue, and the country's demographic advantage, a large young workforce entering the labour market, assumes those workers have jobs to enter.

The stakes are national. Outsourcing revenue supports consumption, property markets and household incomes across the economy, and a sustained decline would be a macroeconomic event rather than a sectoral one.

Where AI is being applied:

Financial services and digital payments have grown substantially, with mobile wallets achieving wide adoption and supporting credit assessment for a largely underbanked population.

Business process companies themselves are significant AI deployers, using it internally for quality monitoring, agent assistance and workflow automation.

Agriculture and disaster management are important applications given the country's exposure to typhoons and its agricultural employment, with forecasting and damage assessment having direct humanitarian value.

Filipino language technology matters, with Tagalog and numerous regional languages spoken alongside English.

The workforce question:

The Philippines has a young, growing, English-speaking population, which in a world of aging societies is a genuine asset.

The country also exports labour globally, with overseas workers' remittances a major economic pillar, and healthcare workers in particular in demand internationally.

Technical education produces engineering graduates, though advanced AI expertise is limited and senior technical talent is scarce.

Constraints:

Infrastructure is a real problem: internet connectivity has improved but remains inconsistent and expensive relative to income, electricity costs are among the region's highest, and power reliability varies.

Geography, thousands of islands, makes infrastructure expensive.

Compute capacity is minimal, and data centre investment is growing from a low base.

Natural disaster exposure is severe and affects both infrastructure and economic planning.

The outlook:

The Philippines is the clearest test case of whether an economy built on outsourced human labour can transition when that labour is automated.

The advantages are real: English, a young workforce, and existing relationships with the global companies deploying AI. The strategy of becoming the place where AI systems are trained, evaluated and supervised rather than where calls are answered is coherent. Whether it scales to employ the numbers the current model does is the unresolved question.