The agent-ready enterprise: why data governance is the CIO's next big challenge

AI agents need more than access to your data. They need information they can understand, verify and trust.

Nov 18, 2026

Dirk Jan van der Pol, Head of Commerce, Empathy Lab
Brian Gilmore, Global Head of Commerce GTM, EPAM

The concept of being agentic commerce ready sounds deceptively simple. Make your data accessible. Structure it correctly. Keep it accurate and up to date. But in practice, this can expose years of accumulated weaknesses in enterprise data governance.

AI agents need more than information they can read. They need:

  • Complete, consistent product data they can evaluate.
  • Pricing that is accurate in real time, not refreshed overnight.
  • Sustainability credentials that can be verified, not just claimed.
  • Availability data that reflects actual inventory, not a snapshot from six hours ago.

Most enterprise data architectures weren't built for this. They were designed around human users, with information presented in ways people can interpret, even when details are missing or inconsistent.

The readiness gap is significant

According to a 2026 study by Harvard Business Review Analytic Services, just 7% of organizations surveyed said their data was completely ready for AI adoption. And when AI agents encounter incomplete or inaccurate information, they may not recognize the gaps, potentially making decisions based on data they can't reliably trust.

Browser-based agents add another complication. The data in your human-facing interface doesn't always match what's in your structured content layer. Promotional pricing might appear on a page but not in your product feed. Availability could be updated dynamically on your website but not through your API. Sustainability claims might feature in your brand copy without supporting structured evidence. An agent browsing your website can encounter these inconsistencies, whether you intended it to or not.

For CIOs, these aren't simply technical inconsistencies. They can affect how products are evaluated, how purchases are made and ultimately, whether customers trust the outcome.

“CIOs who treat data governance as a strategic priority, rather than a compliance exercise, can turn agent-readiness into a competitive advantage.”
Brian Gilmore, Global Head of Commerce GTM, EPAM

Content governance: agents can repeat what you publish

 

Data tells the agent what you sell. Content tells it what to believe.

Data governance is only half the picture. The other half is content: the descriptions, claims, policies, FAQs and reviews that agents read, summarize and present to consumers – often without a human checking the result.

Most enterprises govern content with human readers in mind. People can often spot an outdated paragraph or question a vague claim. An AI agent may take that information at face value, repeat it with confidence or use it to inform a decision.

4 gaps every CIO needs to address: 

Which version is true?  

Your CMS, PIM, DAM and commerce platform may each hold a different version of the same information. When your website says one thing and your product feed says another, an agent may encounter both. Someone needs to own the process of resolving those inconsistencies.

Are your claims defensible?

Sustainability, health and performance claims written for brand impact may become part of an agent's answer. Without clear evidence to support them, they can create regulatory and reputational risks.

Can agents understand your policies?


Returns, warranties, shipping and eligibility rules are often buried in lengthy text or PDFs. If agents can't reliably interpret them, they may provide incorrect answers or overlook important conditions.

Can your content be used against you? 


Website copy, reviews and other third-party content can contain malicious instructions designed to manipulate AI agents. This creates a security risk that makes content governance an IT concern as well as a marketing one.

Content needs the same discipline as data: version control, regular updates, clear ownership and traceability. Enterprises need to know which information is current, where it came from and, where possible, what an agent was shown.

Marketing owns what the brand says. The CIO has a critical role in ensuring it is accurate, structured, consistent and safe once a machine is the reader.

“When AI agents become the reader, content governance becomes a business risk, not just a brand responsibility.”
Dirk Jan van der Pol, Head of Commerce, Empathy Lab

When the agent gets it wrong, who's responsible?

 

AI agents are checking. They can read your claims, policies and reviews, repeat inaccurate information and act on it. If that content is stale, contradictory or manipulated, the first person to find out might be a customer, a regulator or a competitor – not your content team. And when that happens, “marketing wrote it” won't be a good enough excuse.

That's why marketing, IT, legal and commerce need to work together. The goal isn't to control every answer an AI agent gives. It's to make sure the information it encounters is accurate, consistent and trustworthy, while protecting the business from the risks of getting it wrong.

But even when your data is ready, there's another question: can you trust what AI agents do with it? Read more in our next article: When AI agents become your customers: a CIO's guide to staying in control.

If you want to see what agentic commerce means for your business, take our free Agentic Commerce Readiness Starter Assessment to see where your gaps are and what to prioritize next.

Contributors in this article

Dirk Jan van der Pol
Head of Commerce, Empathy Lab
Brian Gilmore
Global Head of Commerce GTM, EPAM
Stay up-to-date with everything Empathy Lab
Follow Empathy Lab on LinkedIn