The pitch deck practically wrote itself: ninety-five percent of corporate AI pilots crater without moving the profit needle, according to MIT. So what if you built something different?
That's the wager a Dallas-area startup called CogitoAI put to investors this spring, and it appears to have landed. The company closed an oversubscribed seed round in late May—more than $3.5 million, 40% beyond the target—from a clutch of heavyweight backers including Ross Perot Jr., former Intel chief Pat Gelsinger, and Helmerich & Payne Chairman Hans Helmerich. The timing wasn't accidental. CogitoAI had just started booking its first commercial revenues.
For a company that announced its exit from stealth in early February, the investor roster reads like a Dallas power lunch. Perot chairs Perot Companies and the sprawling Hillwood development empire. Gelsinger, who stepped down from Intel in December 2024, brings Silicon Valley credibility. Helmerich oversees a publicly traded energy services firm. None of them, apparently, were deterred by the fact that CogitoAI's address situation is a bit of a riddle: press materials say Dallas, LinkedIn says London.
The ROI Problem
What they're buying into is a platform that deploys AI agents across business functions—sales, marketing, customer support, HR, finance, engineering—with an insistence on measurable returns rather than science projects. It's enterprise software, but the company is positioning it as a corrective to the breathless experimentation that characterized the first wave of generative AI adoption.
The 2025 MIT study CogitoAI cites has become something of a rallying cry for vendors promising discipline over demo-ware. Whether the startup can deliver where others have stumbled remains an open question, but the early commercial signal seems to have convinced investors there's at least smoke.
CogitoAI points to an initial deployment at what it describes as an international software company—no name disclosed—where the platform reportedly drove cost savings and revenue lifts of 20 to 50 percent across covered processes, according to company figures that have not been independently verified. That rollout spanned 15 use cases, pulled from 10 data sources, and reached the entire employee base through what the company calls a "Master Agent." It's the kind of claim that would typically require a case study with real attribution to hold up under scrutiny, but for a seed-stage company, directional traction is often enough.
A Team Built for Scale
CEO Andy Evans comes with four decades in enterprise software, including senior stints at portfolio companies backed by Blackstone, Apax, and Goldman Sachs, plus an early career tour at McKinsey. In February, the company brought on Gregory W. Slayton as chairman—a Silicon Valley CEO and VC who advised Google and Salesforce before their IPOs. CTO Naveen Kumar arrived in April with 28 years in AI and data systems at IBM, HP, and Conga.
The bench suggests ambition beyond a small-scale services play. So does the hiring: CogitoAI had roughly 18 employees as of late May and has been posting software engineering roles in Ahmedabad, India, a signal it's building for operational leverage.
What's Actually Shipping
Details on the product remain somewhat opaque. The company announced a "next-generation beta" deployed to select clients in the second quarter of this year, per its February launch materials. It lists technology partnerships with Google Cloud, Atlassian, Cloudflare, MongoDB, and Neo4j, though the specifics of those relationships haven't been made public. Standard startup opacity, perhaps, or strategic caution in a crowded market.
CogitoAI hasn't disclosed lead investors, valuation, or customer names beyond the generic software company reference. That's typical for early-stage rounds, but it leaves observers with more questions than answers about what, exactly, separates this platform from the dozens of other enterprise AI plays competing for the same budget dollars.
The fundraise will presumably fuel commercial expansion and product development through the back half of the year. Whether CogitoAI can convert its early momentum into sustained growth—and prove that enterprise AI can, in fact, escape the 95% failure rate—will depend on execution in a market that's already littered with unfulfilled promises.
For now, it has cash, credibility, and a thesis that resonates with a particular kind of investor: the ones who've watched too many AI pilots fizzle. That might be enough.
