The Day the Body Shop Blames the Robot

One of the world’s big IT outsourcers is going to walk onto an earnings call before mid-2027, clear its throat, and say the quiet part out loud: “Our core outsourcing revenue is down year-over-year because AI means clients don’t need as many of our humans.”

Not just the usual wallpaper — “efficiency,” “productivity,” “complex macro environment.” A direct line from “AI automation” to “this specific outsourcing segment shrank.”

My bet: by Q2 2027 earnings season, at least one of the current top-10 global IT services firms — think Accenture, TCS, Infosys, Cognizant, IBM Consulting, Capgemini, Wipro, HCLTech, DXC, Tech Mahindra — will report a year-over-year decline in a clearly labeled core outsourcing line and openly pin part of it on AI cannibalization.

When it happens, you’ll hear a new genre of corporate poetry: the earnings script where management celebrates shrinking revenue as visionary self‑harm.

AI vs. The Pyramid Scheme

The economics are already written. Generative AI and automation slash the human labor needed for the exact stuff global outsourcers built empires on: grinding out code, reconciling ledgers, checking documents, running back‑office processes on repeat.

Adjacent sectors are living the prequel. Tech firms from Atlassian to Amazon are cutting thousands of roles while promising AI will “reduce repetitive work” and “increase velocity.” Snap openly cited AI in its layoff memo and got a share‑price pat on the head. Law firms like Clifford Chance are quietly redrawing their staffing pyramids as AI takes over junior document review.

Labor data says the same thing without the corporate PR filter: automation is hitting junior, routine roles first, while experienced workers get more productive and harder to replace. The bottom of the pyramid is being hollowed out; the top is getting taller and sharper.

That’s a problem if you are, say, an Indian IT services giant whose business model is a human pyramid with billing codes. Your core product is bodies-on-bench. AI’s core value proposition is: fewer bodies, same (or better) outcomes.

The Cannibal Has a PR Strategy

Former Genpact CEO NV “Tiger” Tyagarajan has already given the sector its script. At Newsweek’s AI Impact Forum, he basically told services CEOs: if AI can eat your revenue, you should be the one holding the fork.

“I am going to cannibalize my revenue because I love it,” he said. Translation: take the hit, then sell investors a bigger, shinier story about higher‑margin, AI‑first work.

Boards are primed to accept this trade. They’ve spent the last few years cheering AI‑linked layoffs: thousands of jobs gone, billions in “structural savings,” stock up a few points. Once you’ve convinced the market that firing people is “expanding operating leverage,” it’s a small step to admitting that AI is shrinking a revenue line – as long as you also promise fatter margins and “AI‑managed services” to replace it.

That’s why this won’t look like a confession. It will be a victory lap with a footnote. Picture a Q2 2027 call where a CFO says something like:

“Legacy application maintenance declined low single digits year-over-year, primarily due to client adoption of our AI-enabled delivery models, which reduce FTE requirements. This cannibalization is consistent with our strategy to shift toward higher-margin AI services.”

That sentence is the finish line for this forecast: a clearly named outsourcing segment, negative year-over-year growth, and AI explicitly blamed for reduced demand or spend — wrapped in a story of deliberate self‑disruption.

Why Someone Blinks by 2027

Two forces have been delaying the admission.

First, creative accounting. If you quietly bury “Application Maintenance” and “Traditional BPO” inside a new “Cloud, Data & AI” lump, nobody can prove cannibalization. It’s the corporate equivalent of pushing old leftovers to the back of the fridge and calling the whole shelf “meal prep.”

Second, narrative cowardice. Once you say “AI is shrinking this business,” analysts start asking awkward questions about your total addressable market — and politicians start sniffing around your layoffs.

But the pressures in the opposite direction are now stronger:

• Clients are not idiots. If AI means their finance process needs 40 FTEs instead of 80, they will push that into renewals. “We want AI, and we want the discount.”
• Competitors are selling AI‑light deals. Refuse to embed automation and you lose the work entirely. Keep it, and the absolute ticket size drops.
• Investors increasingly want proof that AI is real, not just a slide. Real means impact on staffing, contracts, and yes, revenue mix.

Somewhere in that top‑10, the math will stop being maskable. A legacy outsourcing segment will tip into negative territory while AI‑branded lines and margins spike. At that moment, the most rational move isn’t to keep blaming “soft demand.” It’s to stand up, salute the damage, and claim you meant to do that.

Who’s Most Likely to Snap First?

The obvious candidates are the firms with the cleanest segmentation and the most breathless AI decks. Accenture, Capgemini, IBM Consulting — the ones forever launching something called “Synapse.ai Cloud Fusion for Outcomes.” They already split out “managed services,” “outsourcing,” “cloud, data & AI” for investors. When a line bends down, everyone can see it.

India‑based giants like TCS, Infosys, Wipro, HCLTech, and Tech Mahindra face the sharpest structural hit — their revenue mix is still heavy on FTE‑priced work — but they also have the strongest incentive to re‑label the old stuff as “AI‑enabled transformation.” Expect more segment musical chairs there.

DXC is the wild card: highly exposed to old‑school infrastructure and apps, already under pressure, and chronically in need of a story about “resetting the portfolio around AI.” If you were going to turn cannibalization into a redemption arc, you’d start from that script.

The first mover will be whichever CEO decides that a one‑quarter revenue bruise is a small price for a decade of being the firm that “saw the AI transition coming.” In other words: whoever is most desperate for a narrative upgrade.

The Future Earnings Script, Leaked

Between now and June 2027, watch for three tells:

Earnings language that finally ties AI to demand, not just delivery — phrases like “AI automation reduced client FTE requirements in our traditional outsourcing offerings.” Segment charts where “Application Management” or “BPO” quietly go negative while “AI & Automation Services” soars. And analyst questions that stop asking if AI is cannibalizing revenue and start asking how much.

When the first big player crosses that line, they’ll insist this is all part of the plan: less low‑margin body‑shopping, more high‑margin AI orchestration, same clients, better profits. They will call it a mix shift, a transformation, a new S‑curve.

You will know it by its simpler name: a billion‑dollar industry finally admitting its product was always hours, and the software just learned to tell time.