Yesterday morning, Cognizant reported third-quarter results that looked, on the surface, like a solid quarter from a solid IT services firm. Revenue of $5.7 billion, up 4.6% year-over-year. Adjusted operating margin of 16.3%, its seventh consecutive quarter of year-over-year expansion. Adjusted EPS of $1.49, up 7.2%. But the number that should interest you, if you run a small business, isn't any of those. It's this: by the end of the quarter, more than half of Cognizant's own software development was AI-assisted, and the company has put 95,000+ frontier-AI certifications in the hands of 45,000+ of its 359,900 employees.
Key points
- Cognizant reported Q3 revenue of $5.7 billion, up 4.6% year-over-year, with its adjusted operating margin expanding for the seventh straight quarter.
- CEO Ravi Kumar S framed the market as moving from AI capability to value realization, with AI acting as a new production function inside client businesses.
- Over 50% of Cognizant's own software development was AI-assisted by quarter-end, and 45,000+ associates hold 95,000+ frontier-AI certifications.
- Project Leap restructuring costs came in lower than feared ($180-250M vs. up to $320M), with roughly $200M in savings redirected into AI investment.
- The enterprise playbook, managed AI running the business with visible proof of work, is now available to small businesses at a flat monthly price instead of an enterprise contract.
Cognizant (Nasdaq: CTSH) beat the midpoint of its own guidance with $5.7 billion in Q3 revenue, up 4.8% in constant currency. GAAP operating margin slipped 60 basis points to 15.4% on restructuring charges, while adjusted operating margin rose 30 basis points to 16.3%, which management said was the highest in four years. GAAP EPS jumped 148% to $1.39; adjusted EPS rose 7.2% to $1.49. The segment picture was uneven in a way worth paying attention to. Financial Services grew 12.3% in constant currency, its third straight double-digit quarter. Communications, Media and Technology grew 10.7%, helped by the Astreya acquisition. Health Sciences fell 0.9% on regulatory and funding pressure, and Products and Resources slipped 0.8% as retail, consumer goods, and hospitality clients pulled back on discretionary spending. Trailing-12-month bookings hit $28.7 billion, up 4%, with a book-to-bill around 1.3x, though Q3 bookings themselves declined 7% year-over-year. Management also flagged Medicaid and Medicare regulatory uncertainty and baked a higher-than-expected Q4 furlough impact into guidance, which was narrowed to $22.05-22.14 billion in full-year revenue. On the capital side, the company generated $925 million in free cash flow for the quarter, bought back $258 million in stock, declared a $0.33 dividend, and has returned $2.3 billion to shareholders year-to-date.
CEO Ravi Kumar S spent the call pushing one framing above all: the market is moving from AI capability to value realization. In plainer terms, companies have stopped being impressed that AI can do things. They now want to know what it produced. That's why the internal adoption numbers matter more than the headline financials. A 360,000-person firm doesn't measure AI-assisted development percentages for a press release. It measures them because that's how the work now gets done. Cognizant is treating AI adoption as a managed, ongoing program with certifications, tracking, and investment behind it, and it cites improving revenue and operating income per employee as the payoff. The company reported 8,500+ AI engagements and described building enterprise AI systems around reimagined workflows. There's also the restructuring angle. Project Leap, the transformation program Cognizant launched in Q2 2026, now carries a total cost estimate of $180-250 million, down from a prior high end of $320 million. Roughly $200 million in in-year savings is being redirected into AI capabilities. When a company that size finds money, it doesn't hand it to shareholders or bank it. It buys AI.
Here's the connection, and it's a real one rather than a stretch. Cognizant's pitch to its clients is essentially: don't build AI yourself, hire a partner who runs it for you and shows you the results. That's the same argument we make to small business owners every day, just at a different price point and a different scale. The giant incumbent and the local HVAC company are converging on the same conclusion from opposite directions. AI works when it's built into the fabric of the business, not bolted on. When Cognizant says AI is a new production function, it means AI should be doing actual work inside the business: answering, drafting, scheduling, invoicing, following up. A small business owner wants the identical thing. The call comes in, the AI drafts the quote, the customer approves it online, the job lands on the calendar, the invoice gets paid, the review request goes out. That loop, closed by AI, is value realization at the scale of a single trades business. The difference is that Cognizant sells this model with enterprise contracts and quarterly bookings in the hundreds of millions. A small business can now buy the same model, a website plus an AI back office that runs it, for one flat monthly fee with no contract. The economics that used to require a Fortune 500 budget have compressed to something a plumber or a salon owner can sign up for on a Tuesday.
The most useful lesson from this earnings call isn't about Cognizant at all. It's about how you should evaluate any AI vendor, including us. Cognizant didn't win seven straight quarters of margin expansion by promising AI. It won them partly by using AI internally and measuring what changed. When you're evaluating AI for your business, apply the same test. Don't ask what the system can do. Ask what it did this week. Which leads came in. Which quotes went out. Which invoices got paid. Which content got published and what it ranked for. A vendor who can't show you the work, with real numbers, on a schedule, is selling capability. A vendor who files a weekly report with rank movers, traffic, and actions taken is selling value realization. That's also why visible AI matters. Cognizant tracks 95,000 certifications because adoption you can't see is adoption you can't manage. A small business owner deserves the same visibility: a live feed of what the AI is doing, confidence scores on drafts, an approval queue, and a human staying in control of anything customer-facing.
One note of balance. Cognizant's Q4 guidance embeds real caution: healthcare funding uncertainty, softer consumer spending, and furlough impacts. Q3 bookings declined 7% year-over-year even with the strong trailing figure. This isn't a market where businesses are spending freely, and that cuts both ways. It's a headwind for anyone selling technology, and it's a tailwind for offerings that replace a stack of per-seat subscriptions with one predictable cost. When budgets tighten, cost-justified systems that produce measurable outcomes win, and tools that merely exist get cut.
Cognizant's quarter is worth reading as a signal, not a stock tip. The largest IT services firms in the world are restructuring themselves around the idea that AI should run the business, and they're measuring everything to prove it. Small business owners don't need 45,000 certified employees or a $200 million investment program to act on the same idea. They need a website that answers, writes, sells, and remembers, with a weekly report showing the work. That model used to be an enterprise luxury. Now it's a flat monthly fee, live in 7 days, and you own everything when you leave. If a 360,000-person company thinks AI belongs inside the operations of the business, your five-person shop deserves the same, and can actually afford it.