The number landed without ceremony. Twenty-five percent. A clean, digestible figure offered by R. Srikrishna, CEO of Hexaware Technologies, during an interview about artificial intelligence's impact on IT project economics. The financial press picked it up, recirculated it, and moved on. But as someone who has spent two decades dissecting the gap between corporate narratives and operational reality, I found myself staring at that number with a different kind of attention.
The data does not lie. But the framing often does.
Let me be precise about what was actually said. The claim, as reported, is that AI integration could reduce IT project costs by 25%. Not will. Could. The modal verb carries more weight than most readers realize. It signals possibility, not commitment. It suggests a range, not a certainty. And it tells us something important about the speaker's strategic intent that the headline missed entirely.
I have audited enough corporate communications over the years to recognize a defensive narrative when I see one. This is not an aggressive bet on AI's transformative potential. This is a pre-emptive concession to a market that has already priced in the disruption. The CEO is not announcing a breakthrough. He is managing expectations.
Here is what the 25% figure actually represents: an acknowledgment that the traditional billable-hour model of IT services is structurally broken, and that the industry's largest cost component — human labor — is about to face its most significant deflationary pressure in decades.
The cost structure logic is sound, on its face. IT services projects allocate 60-80% of their budgets to human resources. Coding, testing, documentation, and basic project management are precisely the areas where generative AI has demonstrated measurable efficiency gains. McKinsey's research on software development productivity suggests a 20-40% reduction in development time for certain task categories. Gartner's projections align with this range. The 25% figure sits comfortably within industry consensus.
But the arithmetic of cost reduction is not the same as the arithmetic of profit preservation. This is where the analysis gets interesting, and where most commentary on this story has stopped short.
Consider the pricing mechanics. If AI reduces the hours required to deliver a project by 25%, and the client is billed on a time-and-materials basis, then revenue falls by 25% alongside costs. The margin remains constant, but the absolute profit shrinks. The company does more work with less compensation. The client captures the entire efficiency gain.
This is the trap that Hexaware's CEO is dancing around with careful language. He mentioned balancing productivity improvements with pricing strategies to maintain profitability. That sentence deserves far more scrutiny than it received. It is an admission that the company expects clients to demand a share of the AI dividend — and that the negotiation over who captures the value from AI-driven efficiency has already begun.
The client is not the only party with leverage. The competitive dynamics are shifting beneath the feet of every mid-tier IT services firm. When Infosys launched its Topaz AI platform, when Wipro committed $1 billion to AI transformation, when Accenture announced $3 billion in AI investments, they were not just building capabilities. They were building narratives that would allow them to justify premium pricing in a market where AI threatens to commoditize their core offerings.
Hexaware, with approximately 30,000 employees and annual revenue in the $1 billion range, occupies an uncomfortable middle position. Too small to match the AI investment war chests of the Tier 1 players. Too large to pivot into a nimble AI-native consultancy. The CEO's public stance on cost reduction serves a dual purpose: it signals to clients that Hexaware understands the new economics, and it signals to investors that the company is not in denial about the structural shift.
The hidden information in this announcement is more revealing than the stated figure. Let me walk through what is not being said.
First, the 25% figure is a concession to client expectations. Every procurement officer in every enterprise IT department has read the same McKinsey reports. They know AI can reduce development costs. They know their current vendors are sitting on efficiency gains. The question is no longer whether prices will fall — it is who will capture the value. By publicly acknowledging the 25% figure, Hexaware's CEO is essentially pre-empting the negotiation. He is saying: we know you know, and we are willing to discuss terms before you force the issue by going to a competitor.
Second, the profitability implication is not straightforwardly positive. If costs fall 25% and prices fall 25%, margins stay flat. If costs fall 25% and prices fall 15%, margins improve. If costs fall 25% and prices fall 35%, margins deteriorate. The CEO's emphasis on "maintaining profitability" suggests he expects pricing pressure to be significant. He is managing the narrative downward, preparing stakeholders for a period of margin compression while the industry figures out new pricing models.
Third, there is a labor market signal embedded in this announcement. A 25% cost reduction in IT services does not happen without a corresponding reduction in headcount requirements. The industry has historically been measured in "billable headcount" — the number of engineers you can deploy on client projects. AI changes this calculus fundamentally. If the same output requires 25% fewer person-hours, the industry needs fewer entry-level programmers. The impact on India's IT employment structure, where companies like Hexaware are headquartered, will be profound.
I have seen this pattern before. In 2017, during the ICO boom, I audited whitepapers and smart contracts for a dozen projects. The ones with flawed tokenomics had something in common: they were excellent at narrative and terrible at math. The ones that survived understood that the equations had to work before the story could be told. The same principle applies here. The 25% figure is a story. The question is whether the underlying economics actually work.
The industry-wide implications extend far beyond Hexaware's individual situation. The global IT services market generates approximately $1.2 trillion in annual revenue. India's IT services sector employs roughly 5 million people across companies like TCS, Infosys, Wipro, HCL, and Tech Mahindra. If AI delivers even half of the efficiency gains that Hexaware's CEO is projecting, the industry will release between $120 billion and $240 billion in annual value. Some of that will flow to clients through lower prices. Some will flow to AI infrastructure providers like OpenAI, Anthropic, and Microsoft. Some will flow to shareholders if companies manage the transition effectively. But a significant portion will come out of the pockets of the industry's workforce.
IDC projects that AI will replace 15-20% of entry-level programming jobs globally by 2028. For India's IT sector, which has built its competitive advantage on low-cost labor arbitrage, this is an existential threat. The industry's entire value proposition — access to a large, skilled, relatively inexpensive workforce — is being eroded by tools that can generate code faster than a junior developer can type it.
The business model is shifting from "time and materials" to "outcome-based." This is not a subtle change in contracting language. It is a fundamental reallocation of risk and reward. Under the traditional model, the client bears the risk of inefficiency — they pay for hours regardless of output. Under an outcome-based model, the vendor bears the risk — they get paid for results, not effort. AI makes outcome-based contracting viable because it reduces the cost of delivering results. But it also exposes vendors to downside risk if their AI-enabled delivery fails to meet expectations.
Here is where my contrarian lens focuses. The prevailing narrative is that AI is a tailwind for IT services companies — a way to improve margins and deliver more value. I think the opposite is more likely true. AI is a deflationary force on the industry's core product. It commoditizes the very skills that have been the industry's primary source of differentiation.
The companies that survive this transition will not be the ones with the best AI tools. They will be the ones that figure out how to monetize outcomes rather than hours. But this is easier said than done. Outcome-based pricing requires a level of trust and measurement that the industry has not yet developed. How do you measure the quality of a software deliverable? How do you account for the long-term maintenance costs that are hidden in every line of code? These are not theoretical questions. They are the practical obstacles to the new pricing paradigm.
The competitive dynamics are worth examining in detail. Hexaware sits in the second tier of Indian IT services companies, competing with the likes of LTIMindtree, Persistent Systems, and Mphasis. These firms have traditionally competed on cost efficiency and specialized vertical expertise. AI threatens to compress the cost differential that has been their primary competitive advantage.
If Accenture can deploy AI to reduce its delivery costs by 25%, and Hexaware can only achieve 15%, the gap between them widens. The Tier 1 firms have the scale to invest in proprietary AI tools, train their workforce at scale, and absorb the transition costs. The mid-tier firms are caught in a bind: they cannot afford to underinvest in AI, but they also cannot afford to overinvest without sacrificing near-term profitability.
This creates a scenario where the mid-tier is squeezed from both directions. From above, the Tier 1 firms with deeper pockets and more advanced AI capabilities. From below, AI-native companies that are building tools to replace entire categories of IT services work. Consider the rise of AI coding assistants that can generate production-ready code from natural language descriptions. If these tools mature to the point where a small team of senior engineers can deliver what used to require a team of fifty, the value proposition of large IT services firms collapses.
There is also a talent war brewing beneath the surface. AI transformation requires AI experts, and those experts are scarce. The Tier 1 firms are aggressively recruiting AI talent, offering premium compensation packages. The AI-native companies are offering equity and the promise of building something new. The mid-tier firms are left with a difficult choice: pay up for scarce talent or risk falling further behind.
The investment implications of this story are more nuanced than the market's initial reaction suggests. When IT services companies announce AI initiatives, their stock prices often tick upward. Investors interpret the announcement as evidence of forward-thinking management. But the market's patience with AI narratives is limited. The question investors should be asking is not whether Hexaware's CEO understands the importance of AI. It is whether the company can convert that understanding into financial performance.
Let me walk through the math. The industry's average operating margin is roughly 20-25% for Indian IT services companies. If AI drives a 25% reduction in delivery costs, the potential margin improvement is significant — if the company can retain even half of that efficiency gain. But the countervailing pressure is pricing. Clients will demand lower prices. The competitive dynamic will force price concessions. The net effect on margins is uncertain.
There is also a timing consideration. AI adoption is not instantaneous. It requires investment in tools, training, and process redesign. The near-term effect of AI investment is likely to be margin compression as companies spend money to build capabilities that will only pay off in the medium term. This is the classic innovator's dilemma: the companies that move too aggressively may destroy their current business before the new one is ready. The companies that move too slowly may be rendered obsolete.
I have seen this pattern play out in other industries. In the early days of cloud computing, the traditional IT infrastructure companies that embraced cloud too aggressively saw their legacy revenue decline faster than their cloud revenue grew. The ones that survived — companies like Microsoft, which pivoted successfully from on-premise software to cloud services — were the ones that managed the transition carefully, protecting their legacy business while building the new one.
The same discipline will be required in IT services. The companies that survive the AI transition will be those that find the optimal pace of change — fast enough to remain relevant, slow enough to protect the cash flows that fund the transformation.
Let me turn to what I believe is the most underappreciated aspect of this story: the changing balance of power between IT services vendors and their clients. The traditional relationship was asymmetric. The vendor had the technical expertise; the client had the money and the need. The vendor's expertise created a barrier to switching. If you wanted to change your IT services provider, you had to transfer knowledge, rebuild infrastructure, and risk disruption.
AI changes this calculus. When a client can use AI tools to prototype applications, generate code, and even deploy production systems, their dependence on external vendors diminishes. The client becomes less captive. The switching costs fall. And the vendor's ability to command premium pricing erodes.
This is why I believe the 25% figure is a strategic communication, not a technical projection. It is Hexaware's way of saying to the market: we understand that the old model is over, and we are ready to negotiate on the new terms.
But here is the uncomfortable question that no one in the industry wants to answer: what happens when the cost reduction exceeds the price reduction? What happens when AI capabilities improve to the point where a 25% cost reduction becomes 50%? The industry's business model was built on a certain volume of human labor. If that labor requirement shrinks faster than the industry can adapt, the result will be a period of significant dislocation.
The employment implications are not just about numbers. They are about the structure of the global workforce. For two decades, the IT services industry has been a path to the middle class for millions of workers in India, the Philippines, and other developing economies. The industry has been a powerful engine of social mobility. AI threatens to dismantle this engine at exactly the moment when the global economy needs more inclusive growth, not less.
There is an irony in the fact that the industry's success was built on efficiency, and now efficiency is the weapon being used against it. The very forces that made Indian IT services companies competitive — a large, educated, English-speaking workforce with strong technical skills — are the forces being commoditized by AI.
The question of whether 25% is the right number is almost beside the point. The number could be 15% or 35%. What matters is the direction of travel. The industry is moving toward a lower-cost, AI-enabled delivery model. The only questions are how fast, how far, and who benefits.
I have been analyzing the intersection of technology and markets for two decades. I have seen narratives come and go. I have watched companies rise and fall on the strength of their storytelling. The ones that survived were the ones that understood the difference between a story and a strategy. The ones that thrived were the ones that built their strategy on a foundation of verifiable data.
In this case, the data is still emerging. We do not yet have reliable statistics on the actual cost reductions achieved by AI-enabled IT services delivery. We have projections and case studies, but the industry-wide evidence base is thin. The honest answer to whether 25% is achievable is: it depends. It depends on the type of project, the maturity of the AI tools, the skill of the delivery team, and the willingness of clients to adopt new working methods.
What I can say with confidence is that the cost structure of IT services is going to change. The industry's response to that change will determine which companies survive and which become cautionary tales. The leaders will be those that embrace outcome-based pricing, invest in AI capabilities, and manage the transition carefully. The laggards will be those that cling to the billable-hour model and hope the disruption passes.
Let me close with a forward-looking observation. The 25% figure will be tested. Not in boardrooms or press releases, but in actual projects with actual budgets and actual timelines. The companies that can demonstrate real cost reductions will build credibility. The companies that cannot will lose it.
I would be watching several signals over the coming quarters. First, whether Hexaware or its peers launch AI-specific platforms or products similar to Infosys's Topaz. Second, whether quarterly earnings reports begin to disclose AI-related efficiency metrics. Third, whether client contracts start shifting from time-and-materials to outcome-based pricing. These signals will tell us whether the industry is actually transforming or merely talking about transformation.
The IT services industry is facing its most significant structural challenge in a generation. The outcome of this challenge will be determined not by the eloquence of CEOs but by the discipline of execution. Ledgers do not lie, only the narrative does. The 25% figure is a narrative. The question is whether the ledgers will support it.
Volatility reveals character, not just value. The same is true for industries. The coming years will reveal which IT services companies have real strategic depth and which were merely riding a wave. Trust the math, ignore the hype. The math will tell you which companies are building durable value and which are constructing elaborate narratives to mask structural decline.
Every major technological shift creates winners and losers. The IT services industry is no exception. The winners will be those that understand that AI is not a tool to be added to the existing model, but a force that requires a fundamentally new model. The losers will be those that treat AI as a marginal improvement to a business model whose time has passed.
I am watching this transformation with a mixture of professional interest and personal concern. Professional interest, because the data will be fascinating. Personal concern, because real people's livelihoods are at stake. The industry's ability to navigate this transition without leaving millions of workers behind will be one of the defining tests of our economic system.
For now, the 25% figure stands as a marker — a public acknowledgment that the old ways are ending. What comes next will be determined by how well the industry manages the transition. The data will tell us the truth. It always does.
