At CDO Magazine’s Executive Boardroom Dinner, data leaders moved past AI buzzwords and got specific about context, readiness, and the foundations that make enterprise AI work

Chicago does a particular kind of evening well in October. The air turns crisp, the light over the river turns gold early, and the whole city seems to pick up its pace, as if everyone knows the season for getting things done is now.

That was the backdrop on October 14, when CDO Magazine brought together senior data, analytics, and AI leaders for its Executive Boardroom Dinner at Chicago Cut Steakhouse, where LaSalle Street meets the river. Informatica from Salesforce was proud to sponsor the evening, and I joined Nick Dobbins, Informatica's Senior Vice President of Solution Engineering, at the table. CDO Magazine built a room for candor. There were no slides, no pitches, just a table of leaders willing to talk honestly about what is and isn't working.

There’s something about a good steakhouse. Nobody orders “beef.” You order a ribeye, a filet, a strip. You name the cut, because precision is the point. It turned out to be a fitting setting, because the theme of the night was precision, specifically about one of the most overused words in enterprise AI — context.

Context is critical to enterprise AI. Everyone at the table agreed on that. But context has become a catch-all concept, and catch-all concepts are hard to act on. When we say AI needs context, we’re really talking about several data management capabilities, and each one deserves its own attention:

• Business process logic: how work actually gets done in an organization.
• Semantic definitions: what terms like "customer" really mean.
• Entity resolution: spotting when different records are the same customer, product, or asset.
• Master data: one trusted record for each of those.
• Relationships between business entities: how all those pieces are inter-related.

Whether the subject is financial data, customer records, assets, reference data, or product information, these together are the components of meaningful business context.

The challenge for these chief data officers (CDOs) is that the context is distributed. It lives in different systems, applications, and data platforms, often with inconsistent definitions and varying degrees of trust. Industry analysts acknowledge this as a challenge, too. Just days after our dinner, Gartner® named AI context platforms in its top strategic technology trends for 2027. According to Gartner, “Without a dedicated context layer, organizations face severe operational risks, including overprivileged agents, compliance failures, low trust, and excessive token consumption.”¹

CDOs and chief information officers (CIOs) need a platform that can discover, understand, connect, and govern this context across their data estate, and then make it consistently available to the business use cases, analytics, and AI agents that depend on it. Context isn't simply about giving AI more information. It's about making sure AI has the right business meaning, relationships, and trustworthy data to act accurately.

AI-ready data is no longer just an industry phrase. It’s emerging as a distinct category of enterprise capability and an operational mandate for CDOs. The upside of getting it right is significant. For example, "Gartner predicts that by 2027, organizations that prioritize semantics in AI-ready data will increase agentic AI accuracy by up to 80% and reduce costs by up to 60%."¹ Semantics is only part of the picture, though. What "ready" means in practice depends on the industry.

Around the table, nobody debated whether AI-ready data matters. The debate was over what it takes, and the answer depended on who was talking.

In healthcare, AI-ready data puts more weight on governance, privacy, compliance, and controlled access. Cataloging data isn't enough. Organizations have to make sure the right policies, permissions, and procedures are consistently enforced, including requirements like HIPAA and, where relevant, FedRAMP. That gets more complex across regions with different and evolving regulations.

In financial services, master data management moves to the front. Organizations need authoritative, trusted records of financial entities, consistent business definitions, and a clear view of how those entities relate. And that information has to show up consistently across customer-facing applications, business analytics, reporting, and AI agents.

The priorities differ, but the foundation doesn't: data must be discoverable, accessible, governed, contextual-rich, and trustworthy. And AI-ready data isn't a one-time certification or a generic checklist. It's an ongoing discipline shaped by your industry, regulatory environment, business processes, and AI ambitions.

So what does it take? Around the table, nobody's answer was a single approach. It was a set of capabilities that have to work together:

Find and understand context wherever it lives. That means cataloging data and metadata across every cloud, application, and platform, and capturing what it means to the business, not just its technical structure.

Make it trustworthy and authoritative. Assess and improve data quality first. Then resolve customers, suppliers, products, and financial entities into one trusted record, with the relationships between them intact. That's master data management, and AI agents make it more important, not less.

Govern it where it's used. Policies, access controls, and quality rules have to travel with the data, whether a person or an AI agent is using it.

Make it available safely across the enterprise. Analytics, applications, and agents should all draw on the same governed context instead of each building its own version.

Together, these deliver three outcomes. Trust: you can't govern what you can't see, so agents and data are governed in one place, the data is valid, complete, current, and meaningful, and policies are enforced wherever it's used. Speed: customers get running faster without stitching tools together. Lower cost: fewer overlapping tools, and reduced context stuffing or hallucination loops from poorly defined data.

Informatica has worked on this problem for more than 30 years, long before anyone called it AI-ready data. What's changed is who is using the data. Companies are pivoting to agentic enterprise operations where agents plan, reason, and execute autonomously. As part of Salesforce, we're bringing that foundation together with Data 360, MuleSoft, and Agentforce, so that trusted context reaches agents where they act.

AI-ready data starts with an organization's definitions, processes, and priorities. The right platform makes that discipline sustainable at enterprise scale.

Sitting a few steps from the Chicago River, it was hard to miss the parallel. More than a century ago, the city reversed the river's flow so it would serve the people who depended on it. CDOs are doing similar work now, redirecting data so it flows, in context, to wherever the business and its AI agents need it.

The next day, the same conversation followed us into a much bigger room at Connect:AI Chicago, our first event co-hosted with MuleSoft. The message sounded familiar from a different angle: scaling the agentic enterprise takes data governance, reliable agent actions, and a single control plane for AI, with a data foundation underneath it all.

Thank you to CDO Magazine and to every data leader who pulled up a chair. I left with a lot to think about, and I suspect I wasn't the only one.

How do you get the right context to the AI that needs it? That's the question behind Informatica's AI-ready data intelligence platform, which brings together the catalog, governance, and master data capabilities that context depends on. See how we're approaching it: https://www.informatica.com.