The most telling moment from our gathering last month of 75 senior technology executives — drawn from Fortune 500 companies, enterprise tech giants, high-growth startups, and AI-native market makers — wasn’t about ambitious rollouts or transformation roadmaps. It was a single question many of these leaders said their own enterprise customers keep asking: “I know we need to do AI. How best to proceed?”
That question is a warning signal for every founder selling into enterprise right now.
Yes, we also heard about lean GTM teams armed with agents, experimentation versus compliance, microteams with lightning-fast dev cycles, merging functions, and reorgs designed to accelerate companies in the age of AI. The bleeding edge is moving fast. But the customer base often isn’t.
That gap — between how fast AI-native startups build and how slowly enterprises can absorb, let alone implement, what they’re building — is one of the most consequential dynamics in enterprise sales right now. If you’re a founder, understanding it could be the difference between closing deals and burning runway.
To understand the gap in real-world terms, we surveyed 123 senior operators across every major enterprise function for our inaugural State of AI Transformation report.
These are CEOs, C-Suite executives, and VPs with a median 22 years of operating experience, real purchasing authority, and hands-on implementation responsibility. What they told us should make every enterprise-focused founder rethink their approach.
The Enterprise Has Decided. Now What?
AI has moved firmly into continuous experimentation mode. 77% of respondents are actively executing on AI initiatives, and 21% describe themselves as AI-native. In many cases, experimentation is now a top-down mandate — while others contend with bottom-up tool sprawl. As Kieran Snyder, Microsoft’s VP of AI Transformation, writes in the report’s foreword: “It’s an anarchist’s moment.” But almost no one said they’re still just exploring. The enterprise has decided that AI matters. That part is settled.
The Single Greatest Obstacle: Time
One-third of respondents named the lack of capacity to research and test new tools as their primary obstacle. They describe “an abundance of options” with “similar messaging.” They say they “don’t have bandwidth to test every option out there.” And the fragmentation is real: 69% of the tools named in our survey were cited only once — confirming that the market for enterprise AI tools has become overwhelming relative to the capacity of organizations to evaluate them.
What Founders Get Wrong About Enterprise Buyers
The implications for founders: The public markets have punished SaaS companies as investors reckon with a world where platform AI from Anthropic, OpenAI, and Google can absorb capabilities that used to justify standalone products. Valuations have cratered. The conversation around the “SaaSacre” is loud and impacts the market daily~~, whether or not it’s overblown~~. For many founders building in this environment, the instinct is to move faster, ship more features, and differentiate on technical sophistication.
Our survey suggests that’s the wrong instinct.
What Enterprise Buyers Actually Want
Enterprise operators we surveyed aren’t asking for smarter models or more features. They’re asking for three things:
- Tool connectivity. They want tools that plug into existing systems — HR, CRM, product analytics, communications — and synthesize data across fragmented environments. The most-cited request: “a single pane of glass across all my existing data sources.”
- AI that takes initiative. Operators want AI that takes action autonomously, executing multi-step workflows end to end, not tools that surface recommendations and wait for a human. As respondents put it: tools failed “because they required too much pull and were not proactive enough.”
- Deep domain expertise. General-purpose AI is now table stakes. Operators expect specialization in specific functions — sales, recruiting, finance, legal — and differentiation at the workflow level.
The ROI Measurement Gap — and the Opportunity Inside It
When we asked operators how they measure AI’s impact, roughly 70% told us they don’t. No KPIs. No measurement framework. Many acknowledge they’re estimating productivity gains, guessing at ROI. “We estimate 10% productivity improvement, but it’s difficult to measure” is a common refrain. Where concrete measurement does exist, it shows up in customer-facing or revenue-generating workflows — deflecting 38% of support tickets or reducing cost of sale by 15%.
This is both a problem and an opportunity. If your enterprise buyer can’t measure the value of the AI tools they already have, they’re going to struggle to justify buying yours. Products that instrument their own impact — surfacing before-and-after metrics, time savings, or output quality data — give internal champions something concrete when budget conversations get hard. That kind of measurement infrastructure is a retention mechanism as much as it is a sales tool.
The Fundamentals Haven’t Changed
What struck me most about the findings is how much of successful enterprise selling still comes down to fundamentals that have held true for decades. Trusted referrals still open doors. Deep workflow integration still drives stickiness. Internal champions still determine whether a tool survives the first renewal cycle.
The buying process — and the human dynamics that at this point still come within it — has changed far less in the last few decades.
The operators we surveyed describe AI through an intern analogy: capable, but requiring oversight. They have near-zero tolerance for errors in areas like finance, legal, and compliance. They worry about data leakage~~, and context remains an issue for data. They want to see that a product works on their actual data — messy and distributed as it is — before they commit.Loyalty is scarce: even with tools they use daily, many question whether they’ll renew.
The message from this survey is clear: the founders who win in enterprise AI will be the ones who meet buyers where they are — still figuring out what they need — and treat that uncertainty as an opportunity, not an obstacle. Educate, build trust, show the path. The solutions that stick will be the ones that prove real value inside real workflows, not the ones that shipped the most features.
The enterprise is all-in on AI. The opportunity for founders is in helping leaders figure out how.
The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.
