As AI grows, legacy will compress: McKinsey & Company's Noshir Kaka

McKinsey's Noshir Kaka says AI disruption could unlock a $15 trillion opportunity, but IT services firms must overhaul business models, talent, and pricing to win

The information-technology (IT) services industry has a long way to go, says Noshir Kaka, senior partner, McKinsey & Company. Rather, he says that disruption led by artificial intelligence (AI) will create a $15 trillion opportunity. In an interview with Shivani Shinde on the sidelines of the Nasscom Technology Leadership Forum, he talks of the shift for IT tech services. Edited excerpts.

 

You are saying the AI-led opportunity for IT services is $200 billion-300 billion. Where is this opportunity coming from?

 

When we talk about tech services, the addressable market is roughly $3 trillion globally — insourced, outsourced, onshore, and offshore — everything combined. Of that, we address about $280 billion. Next to that is the software market, which is another $2 trillion. With AI-led services, we can now also go after that segment. So, $3 trillion plus $2 trillion gives us a $5 trillion combined opportunity.

 

Now look beyond tech. Consider all professional services globally — doctors, lawyers, consultants, investment bankers, and journalists — any high-skilled, domain-intensive job. In each of these professions, 30-40 per cent of the tasks are repetitive or administrative, and can potentially be automated.

 

If AI can automate 20-40 per cent of these, that represents another $15 trillion opportunity.

 

So, from $3 trillion in tech services, to $5 trillion including software, and another $15 trillion from automating parts of high-skilled jobs, the new addressable market for tech services is $300 billion-400 billion.

 

What are the challenges for the industry to survive — and grow at double digits — while tapping into this opportunity? What changes in business models are required?

 

There are four or five major shifts we have to navigate. First, we will see compression in the core. We saw this during the digital transition. As AI grows, legacy will compress.

 

Second, this is a fundamental operating model shift. If we move towards platform-led, product-centric, or componentised-delivery models, that is different from a traditional services-plus model. Partnership structures, investment in intellectual protocol, and even organisational design need to change. A product- or platform-backed company is structured differently from a pure services company.

 

Third, commercial models will shift. Enterprise customers are increasingly asking to move from activity-based billing to outcome-based models — whether technical or business outcomes. Many ask for it. But when they see what true outcome-based models require — risk-sharing, accountability, and measurement complexity — they sometimes step back. Still, that shift is inevitable.

 

Fourth, talent and workforce reskilling will be massive. This is not incremental upskilling — it’s a structural shift in skills and roles. Fifth, the “net new” opportunity we discussed is real, but it’s not immediately adjacent to what we’ve traditionally done. It requires building new capabilities, relationships, and domain depth.

 

So yes, services players are well positioned. But it’s not a given. It’s not a birthright. Successfully navigating these five structural shifts will determine who captures that $300 billion-400 billion opportunity.

 

For over a decade we have heard that the industry must move to an outcome-based billing structure. How significant will this shift be in capturing the opportunity you’re describing?

 

If you look at enterprise tech budgets — and this is based on a December 2025 survey of 610 chief information officers (CIOs) — 72 per cent said they planned to increase their tech budgets, on average by 6-8 per cent. Overall spending is not shrinking. 

 

This spending is largely directed toward capturing digital and technology benefits, modernising the tech debt, upgrading infrastructure, and investing in data and AI. You cannot modernise tech debt by simply deploying a large-language model (LLM). It requires structured intervention.

 

But here’s where compression comes in. If tech budgets are rising 6-8 per cent, about 30 per cent of that incremental spend may go into net-new areas like data and AI. The remaining 70 per cent is often repurposed from existing budgets. That means CIOs are cutting in traditional areas like application development and maintenance (ADM) and other legacy services.

 

If I’m a CIO trying to extract savings, I have three broad levers: One, bundle expenditure and renegotiate pricing; two, demand productivity gains — “use AI, but deliver me the same or better output at lower cost”; and three, and move to genuine outcome-based contracts.

 

In theory, outcome-based models sound ideal. But in practice, both sides hesitate. That’s why, despite a decade of discussion, true large-scale outcome-based contracting is still limited. The shift will happen — but it will be gradual and likely concentrated in specific-use cases rather than across the board.

 

There will be a section of people who may genuinely not get jobs because of what they currently know. What happens to them?

 

The reality is that every one of us will be pushed to learn and relearn faster than ever before. Some people will rise to that challenge. Some may not. But the opportunity itself is massive. We are talking about moving from a $3 trillion tech services market to potentially $15 trillion-17 trillion when you include adjacencies across broader services. The market exists. The opportunity exists. The industry is well placed. What will make the difference is whether individuals step up — and whether enterprises do.