Three shifts, one strategy

Management framed the AI transition around three shifts: work is moving from AI assisting employees to AI doing more of it, with humans left to handle exceptions and judgement; value is migrating from generic AI tools to a company's own data, context and decisions; and pricing is moving from billing for effort to charging for outcomes. Coforge offered live proof that the model already pays.

The autonomy stack Coforge is building

The report's title captures the strategy: strengthen the data foundation and the execution layer, and put an intelligence platform between them. Coforge's offering stacks into three tiers.

Data remains the critical layer beneath agentic AI: existing lakes and warehouses are insufficient unless data can be made contextual and usable by agents.

How far enterprises really are

The gap between experimenting with AI and running it autonomously is the whole opportunity. Of Coforge's 200-300 clients, most have started, but very few have reached true autonomy — the narrowing tiers below tell the story.

Three gaps — resiliency, security and control — must close for autonomy to scale, and Coforge is addressing resiliency through EvolveOps, an architecture of 28 agents spanning sensing, detection and action. It has already migrated about 10,000 workloads for a Latin American telecom operator in under six months and automated infrastructure management across seven markets for a European bank; agentic operations now span 160-plus clients.

The delivery engine and the economics of outcomes

The forward-deployed engineering model is the engine of AI-led delivery. New cohorts of 20-25 engineers pass through a 12-week academy every six weeks, and lessons from each deployment feed back into reusable assets — making each subsequent build faster and cheaper, and steadily shifting work from humans to agents. Every one of the top 50 clients now has an autonomy roadmap.

The economics are moving with the technology. Open-weight models now make up about 35 percent of token consumption, from 11-12 percent in early 2025, and the metric that matters is cost per unit of work rather than raw tokens — token costs have fallen roughly 100-fold in seven months even as consumption has risen about 1,000-fold. Beyond that, Coforge's Aernova platform targets autonomous airline retailing, which management believes can add around Rs 620 of revenue per passenger over a decade-long adoption cycle in which more than 80 percent of the travel ecosystem has yet to begin.

The numbers behind the call

Earnings are set to step up sharply, aided by scale and the Encora integration, with margins widening and the multiple compressing quickly as profit compounds.

Sales are modelled to leap from Rs 16,400 crore in FY26 to about Rs 30,060 crore by FY28E, with adjusted earnings per share climbing from Rs 40 to Rs 76. That growth drags the price-earnings multiple down from about 51 times to under 24 times on FY28E, against a target set at 29 times — the gap the brokerage is playing for. Returns on equity ease as a larger equity base is put to work, a point worth noting.

The target for investors

Motilal Oswal reiterates a BUY and keeps Coforge as its top pick, valuing it at 29 times FY28E earnings for a target of Rs 2,200, roughly 23 percent above the current price. The bull case rests on strong deal wins, improving cash conversion and further margin upside from the Encora integration, layered on the AI-led delivery shift. For investors, this is a conviction hold on the growth leader in the mid-tier IT pack; with the shares mid-way in a wide 52-week band, the levers to watch are the pace of forward-deployed-engineer scaling and how quickly clients embrace outcome-based pricing.

Key monitorables and risks

Editor's note: this strategy-session update frames its concerns as “key monitorables” rather than a formal risk list; the points above combine those monitorables with standard sector risks, as distilled by TopNews.