BPO Market’s $435 Billion Expansion Amid AI Automation Threats

Two stories about AI and outsourcing are colliding in 2026. Somehow, both are true.

The first one goes like this: AI is tearing the BPO industry apart from the inside. VC-backed startups are building AI agents for customer service, invoice processing, and claims management — no humans required. A16z calls it the unbundling of BPO: a $300 billion market in 2024, headed toward $525 billion, but with AI-native firms picking apart every service line from the bottom up [1].

Now look at the Philippines. Call center employment there nearly doubled between 2016 and 2025 [2]. Unemployment dropped from 9 percent to about 4 percent over that same stretch.

India still employs around 5 million people in outsourcing. The humans aren’t going anywhere. The workforce keeps growing.

The Collision of Two Forces

The reason both stories are true comes down to a 160-year-old idea called Jevons Paradox. Back in 1865, economist William Stanley Jevons noticed that more efficient coal engines didn’t reduce coal consumption. They increased it — because cheaper energy meant people found more uses for it.

Apollo chief economist Torsten Slok applies the same logic to AI and call centers. When AI makes customer service cheaper and faster, companies just buy more of it [2]. The numbers back him up.

The global BPO market is projected to hit $434.99 billion in 2026 and $491.15 billion by 2030, with customer service alone growing at 11.2 percent annually [3]. Eighty-three percent of executives already use AI in outsourced services, according to Deloitte’s Global Outsourcing Survey [3]. But only 25 percent report measurable cost or service improvements.

The tech is everywhere. The ROI is still hit or miss.

“When AI makes customer service cheaper and faster, companies just buy more of it.”

How AI Is Eating the BPO From Within

The a16z analysis splits BPO work into three layers that AI is chipping away at. The first is front-office customer experience — over $100 billion of total BPO spend. Companies like Decagon now hit 80 percent first-contact resolution with AI agents that handle calls humans used to take [1].

Voice AI quality is good enough now that callers barely notice the difference. The second layer is back-office operations. In healthcare revenue cycle management, startup Camber used generative AI to cut first-submission denials by 80 percent and slash billing processing time in half [1].

In logistics, Loop automated invoice reconciliation and claims management — work that used to require teams of human reviewers at traditional freight audit firms. The third layer is application development. Coding assistants like Cursor and AI web app builders are shrinking demand for outsourced engineering teams.

In-house developers with AI can produce what once needed an offshore team, and non-technical staff can now build internal tools without writing a line of code.

The Startup Advantage and the Incumbent Trap

Legacy BPO companies — Wipro, Infosys, Accenture, Genpact — have a structural problem. Their model charges by the hour and marks up the cost of human labor. Switching to AI-native delivery would eat their margins and undermine their own revenue [1].

A16z describes it as an almost impossible pivot: rebuilding a labor-based business into a product-first AI company would kill the cash cows and require a total cultural overhaul. Startups don’t have that problem. They sell outcomes, not hours.

Foundation sells policy administration as a service to insurers, using AI internally to do the work. Long Lake does the same for homeowners associations. These are tech-enabled services — the customer never sees where humans end and AI begins [1].

The pressure is real. Offshore wages in India are rising at mid-to-high single digits every year, slowly killing the cost arbitrage that built the industry [4]. Portage analysts say BPOs must shift to outcome-based pricing or watch business move to in-house global capability centers — or to direct competitors [4].

The Human Side: Growth Amid Disruption

But the Fortune data complicates the disruption narrative. Philippines call center employment nearly doubled as AI advanced. The wage gap is still enormous: $243 to $1,948 per month in the Philippines versus roughly $2,866 in the United States [2].

Stanford economist Erik Brynjolfsson found in a 2023 study that AI assistants boosted agent productivity by 14 percent per hour. Earlier research showed eBay’s AI translation tools increased international exports by 17.5 percent. Higher productivity, in this reading, doesn’t kill jobs — it expands demand.

UVA professor Emma Harrington calls this an old pattern in a new coat. When technology makes cross-border trade cheaper, more trade happens [2]. Language AI lets companies serve more markets in more languages, which creates more demand for the offshore workers who handle the edge cases AI still can’t manage.

The workers themselves are adapting. Many are training for roles in AI oversight, data labeling, and process design — jobs that sit alongside the automation rather than competing with it [2].

The 85 Percent Hypothesis

Some analysts say the disruption is just running ahead of the headlines. APG Technology estimates that 85 percent of currently outsourced work is automatable — the routine, rules-based, high-volume tasks AI can handle end-to-end with humans only reviewing edge cases [5]. That covers the sweet spot of traditional BPO: tech support, helpdesk, mortgage processing, claims administration, finance and accounting.

Generative AI hit contact centers hard in 2023 and 2024, absorbing exactly the kinds of work BPO scaled on. In many organizations, 85 percent of tickets are routine — status checks, data updates, simple eligibility decisions, document intake [5].

Nobody knows how fast this will really happen. Gartner warns that over 40 percent of agentic AI projects could be abandoned by 2027 due to immaturity and unclear ROI [5].

The direction is clear. The pace is anything but.

“AI assistants boosted agent productivity by 14 percent per hour.”

The New Rules of the Game

A Portage report points to a major shift in how BPO gets bought and priced. Traditional contracts paid by the head — full-time equivalents. The emerging model is outcome-based: vendors get paid for transactions completed, service-level targets hit, or business results delivered [4].

This aligns BPO incentives with automation instead of against it. AI agents are now handling cases once reserved for humans — dispute settlement, unauthorized access investigations, first-pass claims resolution [4]. The expectation is that one-third to one-half of all call center operations will be automated within five years.

The question isn’t whether the work changes. It’s whether incumbents or startups capture the value.

The Geography of Disruption

For countries built on outsourcing, the stakes are existential. The Philippines pulls in over $40 billion in annual BPO revenue with nearly 2 million workers. India’s IT services and BPO sector employs roughly 5 million people and makes up a big chunk of services exports [2].

Both are racing to reskill for higher-value work — AI oversight, data labeling, process design [6]. With 83 percent of executives already using AI in outsourced services [3], companies are demanding this from vendors. Organizations choosing outsourcing partners in 2026 increasingly require AI delivery capability, not just cheap labor.

Vendors that can’t show AI-augmented workflows and measurable productivity data are losing enterprise contracts to those that can. Portage says the BPO industry is at an inflection point where generative AI is speeding up trends already underway: wage inflation in traditional hubs, buyer demand for value-added services, and the shift from seat-based to outcome-based pricing [4].

The firms that survive will be those that treat AI not as a cost-cutting tool within an unchanged model, but as the foundation of an entirely new way of delivering work. A $435 billion market, still growing, with offshore employment still expanding. But the nature of that work, who controls it, and how it’s priced are all being rewritten.

Government programs in the Philippines are training agents for AI-related roles. India’s tech sector is investing heavily in upskilling programs [2].

The next five years will decide whether the incumbents pull off the pivot or the startups pick them apart piece by piece.