A number is circulating in Indian business WhatsApp groups right now, usually pasted without context: 56% of CEOs worldwide say AI delivered no measurable financial return last year. It comes from PwC's 2026 Global CEO Survey of 4,454 chief executives across 95 countries, and it's real. It's also being read as an argument against automation, which is the wrong conclusion to draw from it.
Look at how those failures actually happened, and the pattern isn't "AI doesn't work." It's "a broad AI mandate, with no single owner, no defined task, and no way to check whether output improved" — a scoping failure more than a technology one. Only 12% of the CEOs in that same survey reported gains in both cost and revenue. What separated them wasn't a bigger budget. PwC found they were three times more likely to have a defined process and a way to validate results before scaling further.
India's own numbers tell a sharper version of the same story. Google and the India SME Forum surveyed 3,249 MSMEs this year for a report they're calling The Google Dividend, and found that workflow automation and supply chain optimisation could lift profitability for Indian MSMEs by 30-35%. Nearly 60% of the businesses already using AI reported double-digit revenue growth. These aren't enterprise transformation programs. Most are small businesses running one automated process well.
Two ways SMEs approach AI
The difference sits in what gets measured before anything is built. A scoped automation, like an invoicing workflow or a WhatsApp response system, has a clear before-and-after: hours spent per week, errors caught, response time. A broad AI initiative usually doesn't, which is why six months in nobody can say whether it worked. That isn't a failure of the technology. It's a failure to define what "worked" would even look like at the start.
Scoping down is harder than it sounds for an owner already stretched across sales, staffing, and compliance. Picking one task means admitting the other nine won't get touched yet, and it's tempting to want all of it solved at once. The SMEs showing up in the profitability numbers made that trade anyway, and it's the reason the numbers exist at all.
Where the ROI actually shows up
Start with the task that costs your team the most repeated hours in a normal week — invoicing, follow-ups, scheduling, and customer replies are the usual candidates. Price out what a month of that task costs in hours before building anything. That number, not a platform demo, is what actually tells you whether the project is worth doing.