Call Center Jobs Are Shrinking. Is the BPO Industry Ready to Change How It Makes Money?

For most of the modern BPO industry's history, the economics were relatively easy to understand. The debate around AI call center jobs is exposing how quickly those economics are changing.

A client had work that needed to be done. The BPO provided qualified people to perform it. More transactions usually required more employees. More employees produced more billable hours, and more billable hours generated more revenue.

It was an effective model because the amount of human effort required was closely connected to the value being delivered.

AI in BPO is beginning to break that relationship.

AI customer service tools can already summarize conversations, retrieve knowledge, draft responses, classify inquiries, automate repetitive processes, and handle increasingly complex customer-service transactions. A well-designed operation may therefore need fewer people to perform the same volume of work, or the same number of people may be able to accomplish substantially more.

For clients, that creates obvious opportunities.

For BPO companies, it creates a more uncomfortable question:

If technology allows us to deliver better results using significantly less human effort, what exactly should the client be paying us for?

After almost 19 years in the BPO industry, from frontline operations to executive leadership, I believe this question will become much more important than the debate over whether AI will simply “replace call center agents.”

The more fundamental issue is that outsourcing (and much of BPO pricing) was built around monetizing effort. AI is making effort cheaper. The industry's next commercial model therefore needs to become better at monetizing value.

AI and Call Center Jobs: Something Has Clearly Changed

The employment data are beginning to show that this is more than a theoretical discussion.

Revelio Labs reported in October 2026 that global call-center employment had fallen 5.3% from its December 2023 peak, ending more than a decade of sustained growth. Call-center headcount has now declined year over year for eight consecutive quarters. (Source: Revelio Labs, October 2026)

The finding needs to be interpreted carefully.

Revelio does not claim that artificial intelligence is solely responsible. Post-pandemic hiring normalization, corporate cost-cutting, and broader economic conditions may also be contributing. However, researchers found that call-center employment began diverging noticeably from other white-collar occupations after the introduction of generative AI. In the 12 middle-income countries examined, other office employment remained relatively resilient while call-center employment stalled or declined. (Source: Revelio Labs, October 2026)

Even more important is how the contraction is happening.

This is not primarily a story of mass layoffs.

The rate of workers entering call-center employment in the countries studied fell from 25.1 to 17.2 per 100 workers after November 2022. Exit rates also declined, but not as quickly. Companies appear to be hiring fewer people and replacing fewer positions rather than eliminating entire workforces overnight. (Source: Revelio Labs, October 2026)

For the Philippine BPO industry, that distinction matters.

The immediate future may not involve hundreds of thousands of agents suddenly losing their jobs. The transition may be quieter: smaller hiring classes, lower replacement hiring, more productivity expected from every employee, and clients asking why an account that required 100 people five years ago should still require 100 people today.

The Philippine industry's own forecasts already suggest that revenue and headcount are beginning to separate.

IBPAP projects Philippine IT-BPM revenue of between $43.3 billion and $50.5 billion by 2028, with employment between 1.85 million and 2.14 million FTEs. Its outlook explicitly describes the next stage of the industry as a shift from workforce scale toward higher-value capabilities and AI-enabled talent. (Source: IBPAP, 2028 Industry Outlook)

That is a major change.

The Philippine BPO industry can potentially continue growing even when employment does not grow at the same rate.

Which brings us to the commercial problem.

The Traditional BPO Pricing Model Can Punish Productivity

Consider a simplified example.

A client currently needs 100 employees to operate a Customer Service program.

The BPO introduces AI-assisted knowledge retrieval, automated summarization, workflow automation, and better self-service. Productivity increases enough that the operation eventually requires only 80 people to manage comparable volumes and service levels.

It would be inaccurate to say that the provider automatically loses exactly 20% of its revenue. Real outsourcing contracts are more complicated than that. They may contain minimum staffing commitments, management fees, technology charges, volume bands, dedicated infrastructure, and other commercial arrangements.

Nor should we pretend that the cost of the 20 employees simply disappears.

AI has its own cost structure.

Someone has to pay for platform licenses, model consumption, systems integration, data preparation, security, engineering, monitoring, quality assurance, governance and the more specialized employees required to manage AI-assisted operations.

HFS Research makes this point directly: AI is not simply shrinking technology budgets. It is changing who carries the cost. Its 2026 research found AI licenses, tokens, and infrastructure costs growing faster than overall IT spending, while providers are increasingly being asked to absorb some of the volatility associated with those technologies. (Source: HFS Research, 2026)

So the real commercial question is more sophisticated than:

“If we need fewer agents, should the client pay less?”

The real question is:

“When AI creates measurable productivity gains, how should that value be divided between the company buying the service and the provider that helped create the improvement?”

That is where traditional BPO pricing begins to struggle.

If the provider invests in automation, improves productivity and reduces labor requirements, the client will eventually expect part of that efficiency to appear in its pricing.

That is reasonable.

But if every productivity improvement simply results in lower BPO revenue while the provider carries the technology investment and implementation risk, then the commercial structure begins to discourage innovation.

We end up with a strange situation in which a BPO company may be financially rewarded for consuming more labor and financially penalized for becoming more efficient.

The Market Is Already Moving Beyond Pure FTE Pricing

This change is already visible in the global services industry. It is also reshaping outsourcing pricing as buyers look beyond billable staff time.

Reuters reported in August 2026 that major Indian IT-services companies are increasingly moving away from contracts based purely on billable staff time and toward pricing linked to productivity, cost savings and measurable outcomes. Tata Consultancy Services said approximately 80% of contracts in its business-services segment now use performance-based pricing. Clients are simultaneously demanding substantial productivity improvements as AI changes the economics of technology services. (Source: Reuters, August 2026)

But declaring the death of FTE pricing would be premature.

HFS Research found that 82% of organizations surveyed still rely on FTE, SLA or license-based structures when contracting AI-related services. The reason is practical: outcomes remain difficult to baseline, attribute, and measure. The conclusion is that hybrid pricing, rather than purely outcome-based contracts, is likely to dominate the near term. (Source: HFS Research, 2026)

That makes sense to anyone who has actually operated an outsourced account.

Suppose a BPO accepts an outcome-based agreement tied to CSAT.

What happens when the client's product fails?

What happens when logistics delays deliveries?

What happens when the client's pricing policy frustrates customers?

What happens when the CRM goes offline?

The BPO controls some parts of the Customer Experience, but not all of them.

The same problem exists on the other side.

If the provider develops an AI workflow that improves productivity by 25%, who paid for the development? Who maintains it? Who owns the intellectual property? What baseline determines the savings? How long should gain-sharing continue? What happens when volumes change?

The more practical direction is likely to be hybrid commercial models that combine dedicated operational capacity with clearly defined productivity incentives, transaction pricing or gain-sharing where results can be credibly measured.

Traditional FTE pricing will remain appropriate where work requires dedicated capacity, human judgment or significant variability.

Transaction pricing can work where the unit of work is clear.

Outcome-based components can work when results can reasonably be attributed to the provider. That is where an outcome based outsourcing model can be effective when the baseline and accountability are clearly defined.

Gain-sharing can work when both client and provider agree on a baseline and can verify the productivity improvement.

The principle should be straightforward:

Otherwise, the incentives are wrong.

AI Efficiency Is Only Valuable If the Customer Outcome Survives

There is another mistake I believe executives should avoid: assuming that reducing Customer Service headcount automatically proves that AI customer service has succeeded.

It does not.

A company can lower labor cost and still make the customer experience worse.

A chatbot can “contain” a customer who eventually gives up. Average Handle Time can decline while repeat contacts rise. Automation can reduce cost per contact while increasing escalations and customer effort.

Gartner's 2026 research illustrates the balance. Half of customers surveyed said generative AI made Customer Service interactions easier. At the same time, 87% said access to a human representative remained essential when companies use GenAI for service. Customers are not inherently rejecting AI; they are rejecting AI when it becomes an obstacle to resolution. (Source: Gartner, August 2026)

For BPO executives and their clients, productivity therefore needs to be evaluated alongside customer outcomes.

FCR still matters.

CSAT still matters.

Repeat contacts still matter.

Escalations matter.

Customer effort matters.

And so does the full cost to serve, including technology.

A company has not necessarily created value simply because fewer employees appear on the staffing plan.

Real productivity exists when the cost of delivering the service improves without damaging the result the customer or the business receives.

The Workforce Transition Will Be Harder Than “Just Reskill”

The employment side of this transformation also requires more honesty. The workforce effects of generative AI, call center automation and AI-assisted operations will not be evenly distributed.

The usual industry response to AI disruption is to say that workers will be reskilled into higher-value jobs.

That is directionally correct.

It is not automatically true.

Revelio Labs found that some of the occupations surrounding traditional call-center work continue to grow. Software and data employment across the middle-income countries studied increased, as did customer-success, client-relations and technical-support roles. These positions also generally carried higher compensation.

But only 10.8% of observed workers leaving call centers moved into those higher-paying categories. (Source: Revelio Labs, October 2026)

That should force the BPO industry to be more realistic.

An excellent Customer Service representative does not automatically become a data analyst.

A tenured agent does not automatically become a technical-support specialist.

A QA analyst does not automatically become an AI evaluator.

People have different abilities, interests, and learning curves. Higher-value roles also require genuine training investment, not just a two-day workshop and a certificate.

The International Labour Organization estimates that 12.7 million Philippine jobs, more than one quarter of national employment, have some degree of exposure to generative AI. Exposure does not mean elimination; the ILO expects transformation and augmentation to be more significant than outright replacement. But the scale of exposure means businesses cannot afford to treat workforce transition casually. (Source: ILO, February 2026)

Some roles will change.

Some will disappear.

Some people will transition successfully.

Others will not.

That is uncomfortable, but pretending otherwise does not help the employees we say we want to protect.

For Philippine BPO companies, that means identifying roles likely to decline before the decline arrives, assessing which employees have the aptitude for more complex work, building real career pathways, and accepting that future operations may require fewer people, but more capable people.

What This Means for the Philippine BPO Industry

The Philippines should not try to protect every traditional Customer Service task from automation.

That strategy would eventually make us less competitive. The same argument is developed in BPO in Philippines: Why Scale Is No Longer Enough.

Instead, the industry should protect and expand the value of Philippine talent.

That means moving more people toward complex Customer Service, technical support, Customer Success, escalation management, QA, knowledge management, analytics, AI evaluation, workflow supervision, and other work where human judgment and domain expertise continue to matter.

It also means BPO companies must become more sophisticated commercially.

We have spent decades becoming experts at workforce planning, recruiting, training, operations and calculating hourly rates.

The next generation of BPO executives will also need to understand productivity economics, automation investment, AI costs, baselining, gain-sharing and outcome measurement.

That is a different skill set.

Sales teams will need to evolve as well.

The traditional question has been:

“How many agents do you need?”

A better question is increasingly:

“What business outcome are you trying to achieve, and what combination of people, process, and technology will deliver it?”

That changes the relationship from staffing toward an operating partnership.

What This Means for Cebu tele-net

For Cebu tele-net Philippines Corp., this is where Human-Driven. Enhanced by AI. needs to become more than positioning.

It should become an operating principle.

Every process should eventually be evaluated through three lenses: Automate, Augment or Human-Led.

Repetitive tasks with predictable rules should be candidates for automation.

Work where AI can make people faster or better should be augmented.

Interactions involving judgment, empathy, negotiation, complex exceptions, or relationship management should remain deliberately human-led where that creates the better outcome.

But we also need to be commercially realistic.

If CTNP creates measurable productivity for a client, we should be capable of discussing how that value is shared rather than automatically returning to a conversation about how many FTEs are required.

That does not mean abandoning hourly pricing tomorrow.

It means building the measurement, governance, and trust required to offer more sophisticated commercial arrangements where they make sense.

It also means ensuring our operational capability does not fall behind our message.

If we say AI enhances our people, we should be able to demonstrate where.

If we say our people are moving into higher-value work, we should develop actual pathways.

If we tell clients we create business value, we should measure enough of that value to prove it.

That is the harder part.

But it is also the opportunity.

From Selling Effort to Creating Value

The Philippine BPO industry has good reason to be proud of what it has built.

For decades, we created value by providing qualified people, operational discipline, and cost advantages to companies around the world.

That model is not disappearing.

But it is changing.

Artificial intelligence is making some forms of human effort cheaper. It is increasing the productivity of other forms of work. And it is forcing clients to reconsider what they should pay for when a smaller team supported by technology can outperform a larger team operating the traditional way.

The mistake would be to respond by defending headcount for its own sake.

Headcount is not value.

Hours are not value.

Even automation is not value unless it improves an outcome that matters.

The industry's opportunity is to move from simply selling effort toward proving and sharing the value created by that effort.

If BPO companies continue selling only manpower, AI will increasingly look like a competitor.

If we sell expertise, productivity, judgment, and outcomes, AI becomes leverage.

That brings me to the question I believe clients and BPO executives need to start answering together:

If AI allows a BPO to deliver better Customer Service with half the people, should the client pay less because fewer agents are required, or should the provider earn more because it created more value?

I believe the answer is both.

The client should share in the efficiency.

The provider should share in the value it creates.

And the commercial model should give both parties a reason to keep improving.

Because the future of outsourcing will not be determined simply by how many people we employ.

It will be determined by how much value our people, enhanced by technology, can create.