Everyone told you AI was coming for customer service jobs. Companies are spending billions on AI agents, voice bots, and agentic workflows to handle inbound calls and support tickets. So why did offshore call center employment in the Philippines — the world’s largest BPO market — nearly double from 2016 through 2025, reaching 2 million workers?

And why did the country’s unemployment rate fall from 9% in 2021 to around 4% as of March 2026?

A new Fortune report published today digs into this apparent contradiction, and the explanation is one that every business leader deploying AI should understand: the Jevons Paradox.

What the Jevons Paradox Is (and Why It Matters Here)

The Jevons Paradox, first described by 19th-century economist William Stanley Jevons in the context of coal and steam engines, states that when technology makes a resource more efficient to use, total consumption of that resource often goes up rather than down. More efficient coal engines didn’t reduce coal demand — they made coal economical to use in places where it previously wasn’t, which expanded total usage.

The same dynamic appears to be playing out with AI and human labor in customer service.

As AI handles routine queries more efficiently, businesses are able to expand their service operations — offering support at more hours, across more channels, and to more customer segments than they ever could before. That expansion creates new demand for human agents to handle complex escalations, nuanced conversations, and the kinds of interactions where customers still insist on talking to a person.

The result is not fewer call center jobs. It is more of them, but different ones.

What This Means for Business

If you are deploying AI in your customer operations, this research should recalibrate how you think about the outcome.

The goal of AI in customer service was never — or should never have been — eliminating every human touchpoint. The goal is to remove the routine and repetitive work, freeing humans for the conversations that actually require empathy, judgment, or authority to resolve. When AI takes on the simple work, you can afford to offer more contact channels, longer support hours, and faster response times. That broader availability creates more total interactions, not fewer.

This is actually good news for businesses thinking about tools like Omni Voice. A voice AI employee that handles after-hours calls, appointment bookings, and common FAQ queries is not eliminating your team. It is expanding your coverage without expanding headcount — and research suggests customers who get a quick, satisfying first interaction are more likely to escalate to a human when they genuinely need one.

The Jevons effect also explains why companies that have deployed voice AI aggressively are not reporting mass layoffs of customer-facing staff. What they are reporting is better coverage, higher customer satisfaction scores, and staff who spend their time on higher-value conversations instead of answering the same five questions on repeat.

The Skill Shift Is Still Real

None of this means customer service workers are immune to change. The nature of the role is shifting significantly.

The offshore workers thriving in a 2 million-person Philippines BPO market are not doing the same jobs as they were a decade ago. The routine script-reading roles are increasingly handled by AI. The humans who are in demand are those who can manage AI tools, handle complex multi-step cases, and build relationships with customers the AI cannot retain.

This is the argument Enterprise DNA has been making about data and AI skills for years. Automation does not make skills obsolete — it raises the floor of what you need to know to stay relevant. Workers who understand how AI tools work, can interpret data to spot where AI is failing, and can manage AI outputs will be significantly more valuable than those who simply hope the wave passes.

The Philippines BPO growth story is not evidence that AI doesn’t matter. It is evidence that AI is changing the shape of work faster than organisations can restructure around it — and the companies that will win are the ones training their people for the new shape, not waiting to see which jobs disappear.

Practical Takeaways

If you are a business leader making AI workforce decisions right now, here is what this research suggests:

Do not frame AI as headcount reduction. If that is your primary KPI, you will underinvest in the expansion opportunities AI creates. The better framing is coverage expansion and quality improvement.

Invest in AI skills across your customer service team. Workers who can use AI tools effectively, review AI interactions, and handle escalations thoughtfully are the most valuable employees in a Jevons-effect environment.

Monitor total interaction volume, not just AI resolution rate. If AI is driving expanded coverage, total interactions may grow even as AI handles a larger share. That is a success, not a cost problem.

Think beyond voice. The Philippines BPO growth is happening across chat, email, and digital channels. The Jevons effect applies across all of them. Voice AI is one part of a broader picture.

The data is fairly clear at this point: AI is not eliminating customer service work. It is reshaping what that work looks like — and the businesses and workers who adapt fastest will capture the growth that comes from expanded coverage and better service quality.

---

For businesses exploring how voice AI fits into their operations, you can learn more about Omni Voice by Enterprise DNA. For teams looking to build the AI skills that will matter in this environment, explore the Enterprise DNA learning platform.