The Gartner Data and Analytics Summit in London wrapped up this week after three days at ExCeL London, drawing more than 3,000 chief data and analytics officers and data leaders from across Europe and globally. The central message from Gartner analysts was direct: the reason most AI programs will fail has nothing to do with the AI models. It has to do with the data underneath them.

“Data governance will be the single point of failure for organizations’ AI ambitions,” said Anurag Raj, Director Analyst at Gartner, during Day 2 of the event. That is not a nuanced takeaway buried in footnotes. It was the headline message of a summit that tracks enterprise AI adoption across thousands of organizations every year.

What the Research Actually Shows

Sally Parker, Senior Director Analyst at Gartner, opened Day 1 with findings from the firm’s annual CDAO agenda survey. The data was clear: organizations with low data governance maturity are significantly more likely to be among those that fail to realize AI value.

That is a correlation strong enough that Gartner is using it as a primary lens for evaluating AI readiness. It is not a small effect. Companies that have not built reliable data foundations are not just behind on AI. They are actively undermining every AI investment they make before it has a chance to work.

Day 2 went further, looking at what governance needs to look like in an agentic AI environment. Raj introduced a framing worth paying attention to: governance of AI, by AI, and for AI.

Governance of AI means setting the rules that AI systems must follow. Governance by AI means using AI systems to enforce those rules automatically. Governance for AI means structuring your entire data environment in a way that makes AI systems useful rather than unreliable. Most organizations are still working on the first of those three.

Three Scenarios, Not One Forecast

Leinar Ramos, VP Analyst at Gartner, presented three scenarios for how AI development plays out over the coming years. The explicit message was that data and analytics leaders should use scenario planning to periodically recalibrate their AI strategies rather than betting on a single prediction.

That is a mature and useful framing at a time when most commentary on AI defaults to either catastrophism or hype. Gartner’s position, backed by survey data from thousands of organizations, is that the range of outcomes is wide and the right strategic response is to build adaptive capability rather than optimize for one future.

The practical implication of this is that organizations which have built strong data foundations will be better positioned in any of the three scenarios. Data governance is not a bet on one AI outcome. It is the hedge that works across all of them.

Unstructured Data Is the Next Gap

One specific session that stood out was a Day 2 focus on governing unstructured data. As AI systems increasingly need to work with documents, emails, voice recordings, meeting transcripts, and contract text, the governance frameworks built for structured databases do not hold up.

This is a genuine capability gap. Most organizations built their data governance practices around databases and data warehouses. The explosion of unstructured content as AI input is exposing an entirely new set of problems that existing governance tooling was not designed to handle.

Organizations that want to deploy AI agents effectively across knowledge-intensive work — customer support, contract review, internal reporting, meeting summarization — will need governance frameworks that cover unstructured content. Very few organizations have built that yet.

What This Means for Business

The Gartner Data and Analytics Summit is not a vendor conference. It is where thousands of people who actually run data programs compare notes and hear what Gartner’s research is showing across its client base. When Gartner says data governance is the single point of failure for AI ambitions, that language reflects patterns seen across real programs at real organizations.

A few things are worth taking seriously from this week’s summit:

Governance maturity is measurable before AI investment. Organizations can assess where they sit on the governance maturity spectrum before committing significant resources to AI initiatives. The organizations that do this work first will make better decisions and avoid the most expensive failure modes.

Workforce AI proficiency is becoming a baseline expectation. Gartner has projected that 75% of hiring processes will include AI proficiency assessments by 2027. That is not a distant prediction. It is a shift that HR teams and business leaders need to be planning for now.

AI governance is itself becoming an AI job. The idea of governance by AI — using agents to automatically enforce data policies and generate machine-verifiable data contracts — is not theoretical. Gartner has projected that half of organizations will be using this capability by 2030. The teams that understand how data governance works at a deep level will be the ones building and overseeing these systems.

The message from London this week is consistent with what data-mature organizations have been saying for years: AI does not fix bad data. It amplifies it. The organizations that are pulling ahead on AI are the ones that made the foundational investment in data quality and governance first, and are now able to move faster as a result.

If your organization is still in the early stages of building that foundation, the 3,000 CDAOs gathering in London this week would tell you the same thing: this is the highest-leverage work you can do before any other AI investment.

---

Enterprise DNA operates at exactly this intersection. EDNA Learn builds the data skills teams need to understand, interpret, and govern what AI systems are doing. If you want an outside perspective on where your organization sits on the AI readiness spectrum, our Omni Advisory service works with business leaders to build that picture and the roadmap to close the gap.