India and the Philippines: Different Versions of the Same AI Shock
Generative AI is hitting Asia's two biggest outsourcing industries at the same time, but at different points in the value chain.
India built a US$315bn technology sector around software engineering, consulting, and technology services. The Philippines built a $40bn-plus industry heavily weighted toward customer service and business-process work.
AI is attacking labour intensity in both.
But conversations we have had with technology companies suggest something more interesting than a simple retreat from offshoring. Businesses are simultaneously automating work, consolidating it into their strongest offshore centres, and moving those centres up the value chain.
The result may be a paradox. India and the Philippines can continue winning work while employing far fewer incremental people to perform it.
India: AI Meets the Engineering Factory
Indian industry body the National Association of Software and Service Companies (NASSCOM) expects India's technology-services industry to exceed $315bn of revenue in FY26 and employ around six million people. For decades, much of that model depended on a combination of skills, scale, and lower labour costs. Revenue growth generally required more engineers. AI is beginning to break that relationship.
Customers are already pressing suppliers for productivity savings. IT Services company Persistent Systems’ CEO Sandeep Kalra has said clients increasingly want the same work delivered 25% to 30% cheaper and faster. And Tech Mahindra CEO Mohit Joshi has said some competitors are pricing in productivity gains of 70% to 80% over five to seven years.
Those numbers sound heroic until you talk to customers.
One technology company we interviewed told us engineering is already running at roughly three times its previous speed in some areas using AI-assisted development.
The company has several hundred engineers and expects to reduce its overall engineering workforce by about 40% over the next two to three years while consolidating more of its activity into India.
Management expects roughly the same order of cost reduction, but is candid that the benefits come from two different sources: AI productivity and geographic consolidation. It is becoming increasingly difficult to separate one from the other.
A second technology company we interviewed described an even more striking operating example.
Engineering squads that previously required ten people have in cases been reduced to three or four through the use of AI tools. Management describes the productivity improvement as up to 70%, before factoring in any additional savings from offshoring.
This is no longer a theoretical debate being conducted in analyst reports. Companies are redesigning engineering organisations now.
An Awkward Wrinkle in the India Story
There is another irony.
Even before generative AI accelerated, parts of India's technology labour market were becoming expensive. More than one company we interviewed told us that some specialist Indian roles had reached costs above comparable Australian positions. That does not mean India loses. Indeed, the company concerned is consolidating more engineering there. It means India's competitive advantage is changing.
Scale, deep technical skills, infrastructure, ecosystem maturity, and management capability matter increasingly alongside labour cost. In that sense, India is being pushed toward exactly the higher-value position policymakers and industry leaders have wanted for years, only rather more abruptly than they might have preferred.
The hiring model is changing with it.
Former Infosys CFO V. Balakrishnan has argued that coding agents mean the industry no longer needs the enormous intake of entry-level engineers on which its traditional pyramid was built. That is probably directionally right. The more sophisticated question is whether new categories of work expand fast enough to offset AI-driven productivity in existing ones.
Markets Are Already Taking a View
Public markets are not waiting for the answer.
India's Nifty IT index fell close to 30% during the H1 2026 before recovering somewhat and remained around 20% down for the year by late August. TCS has suffered a particularly significant reduction in market value over the past two years.
Some brokerages estimate that 12% to 15% of sector revenue faces direct AI-driven pressure. Those estimates should be treated as what they are: forecasts. They are useful because they tell us what investors fear.
The operator evidence tells us something more concrete: productivity improvements are already large enough to affect staffing, location decisions, and pricing models.
The Philippines: Fewer Low-Value Jobs, but Not Necessarily Less Work
The Philippines presents a different version of the same problem.
Its IT and business-process management industry generated more than $40bn of revenue and employed roughly 1.9 million people in 2025. Industry body the IT and Business Process Association of the Philippines (IBPAP) expects both revenue and employment to grow further in 2026.
The more revealing numbers sit further out. Industry body IBPAP has reduced its 2028 revenue ambition from a previously stated $59bn to a range of $43bn to $50bn and revised its employment target down from 2.5 million workers to a range of 1.85 million to 2.14 million.
The industry does not necessarily shrink. The jobs maths changes.
Basic functionality like lower-level customer enquiries, password resets, data entry, transcription, appointment scheduling, and templated messaging are among the easiest activities to automate. Work requiring empathy, judgement, accountability, or specialist knowledge is more durable. That distinction is visible inside companies already.
Manila May Win Before It Shrinks
One technology company we interviewed has had roughly 500 employees in the Philippines across customer service, finance, IT, and data.
Despite growing automation, management expects that number to remain broadly flat for the next two to three years. The reason is important.
As lower-value activities are automated, the company is simultaneously moving more sophisticated finance and back-office work into Manila from higher-cost locations such as Europe. In other words, Manila can lose tasks while gaining work.
Longer term, the employment outlook becomes less comfortable. The same business is growing revenue strongly but expects its wider support workforce to remain broadly static in the near term and believes that workforce could ultimately become 30% to 40% smaller over five years. That is the decoupling of revenue growth from employment growth in one example.
Manila could still win the next round of offshoring, even while eventually employing fewer people.
The Human Advantage Still Matters
The Philippines also retains an advantage that rarely appears in wage comparisons. One technology company we interviewed told us that English-speaking customers generally respond more positively to Filipino customer-service voices than other nationalities for whom English is not a first language. That is one company's experience, not a universal truth, but it matters.
Accent, cultural familiarity, empathy and customer acceptance still influence where customer-facing work is located. In activities where the interaction itself is part of the service, being cheaper is not enough. This is another reason the impact of AI will differ sharply by task.
Public Markets Are Pricing Uncertainty
TaskUs offers one indication of how investors are thinking about Philippine exposure. The company's market capitalisation fell sharply from its mid-2025 levels into 2026 even while the underlying business continued to grow.
Concentrix, which is considerably more diversified but also has significant Philippine delivery exposure, has also faced repeated investor concern around the impact of generative AI.
This does not prove that either business model is broken. It tells us markets are placing a discount on uncertainty over pricing, margins, and labour intensity well before the long-term operational outcome is clear. That is exactly what markets tend to do.
AI Is Attacking Tasks Faster than Locations
The most important conclusion from the evidence is that AI is attacking tasks faster than it is attacking locations. India can gain engineering work from elsewhere in Asia while using fewer engineers to perform it. Manila can gain finance work from Europe while automating repetitive work already sitting in the Philippines.
Those are not contradictions. They are likely to be features of the transition.
We think the adjustment may occur in three broad stages. First, companies automate the simplest, most repetitive tasks. Second, they consolidate the work that remains into their most capable offshore centres. Third, total headcount compresses as AI penetrates progressively more complex activity.
That leaves India and the Philippines with the same paradox. They may continue to gain global services market share while creating far fewer incremental jobs. The countries can remain winners even when labour is not.
See also: AI is not killing outsourcing. But it is changing the way it gets bought, sold, and priced.