We wrote about “Excel jockeys” back in 2024 — the cost of running a revenue org on manually maintained spreadsheets instead of real systems. The argument then was about error rates and wasted hours. In 2026, the same underlying habit has a new name and a much bigger blast radius: companies are pointing AI at exactly the same ungoverned, manually patched data, except now nothing stops to double-check it before it drives a decision.
Everyone Is Doing AI. Almost No One Says Their Data Is Ready.
Dun & Bradstreet’s AI Momentum Survey, fielded in May 2026 across 10,000 businesses, found that 97% report active AI initiatives. Only 5% say their data is actually ready to support them. CData’s “State of AI Data Connectivity: 2026 Outlook,” a survey of more than 200 enterprise data and AI leaders, puts a similar number on infrastructure specifically: just 6% say their data infrastructure is fully ready for AI, even though 53% of organizations struggling with AI implementations trace the problem directly back to immature data systems.
The Gap Isn’t Awareness. It’s a Confidence Problem With the Numbers to Prove It.
The most useful data point in this year’s research isn’t the low-readiness number — it’s the gap between what leaders claim and what they admit when asked a follow-up question. Precisely and Drexel University’s LeBow College of Business surveyed 505 senior data and analytics leaders (managers through C-suite, at companies with 1,000+ employees or $250 million-plus in revenue) for their 2026 State of Data Integrity and AI Readiness report. Eighty-seven percent claimed AI-ready infrastructure, yet 42% separately named infrastructure as their top obstacle. Eighty-eight percent reported their data was AI-ready, yet 43% named data readiness as their single biggest barrier. That’s not a rounding error — it’s two different answers to the same question, depending on how directly you ask it, and it tracks with what CData’s survey found on the ground: AI teams spend more than a quarter of their time on data preparation instead of the work they were hired to do.
Where the Time and the Trust Actually Go
Dun & Bradstreet’s respondents named the specific blockers behind that 5% readiness number: 50% cite problems simply accessing their own data, 44% cite privacy and compliance risk, 40% cite data quality and integrity, 38% cite system integration, and 37% cite a shortage of qualified people to do the work. Hex’s State of Data Teams 2026 survey found the same theme from inside data teams themselves — data quality and trust concerns were cited by 31% as the single biggest obstacle to AI adoption, nearly twice the rate of any other concern, even as 58% of data teams grew headcount through late 2025. Companies are investing in more people and more AI. They are not, by their own admission, investing enough in the unglamorous work of making the data underneath trustworthy.
Governance Is the Variable That Actually Moves the Needle
The Precisely/Drexel research also isolated what changes the outcome: 71% of leaders report high trust in their data when a formal governance program is in place, versus 50% without one. Sixty-three percent of surveyed organizations now have some form of AI governance program, and 40% have simply extended their existing data governance to cover AI rather than building something new. The catch is that governance without accountability doesn’t close the gap on its own — only 31% of those programs have metrics tied to actual business KPIs, which means most of them can say a policy exists without being able to say whether it’s working.
Our Verdict
The 2024 version of this argument was about spreadsheets: manual data work is slow, error-prone, and expensive. The 2026 version is the same problem wearing an AI badge — 97% adoption against 5% readiness, 87% claimed AI-ready infrastructure against 42% naming infrastructure as the top blocker. The fix hasn’t changed either. It’s a named owner for each data domain, validation before data enters the system rather than after a model acts on it, and governance metrics tied to a business outcome someone is actually accountable for — not just a policy on file. Companies that had already done that work before 2026 are the ones showing up in the 6% and the 5%. Everyone else is running AI on data they wouldn’t sign off on for a board report, just faster than they were in 2024.
If you want a clear read on how AI-ready your revenue data actually is before you point more of it at production systems, that’s a conversation we have with clients regularly.