Every year, the OpenText Content Cloud team heads to the AI+IM Global Summit. Every year, we come back with a clearer picture of what's actually happening inside enterprise organizations—not the version that shows up in analyst reports, but the unfiltered version that surfaces when practitioners are in a room together with no slides and no agenda to protect. 

This year was different. 

The conversation has shifted. Organizations aren't debating whether to invest in AI. They've already invested.  

We ran two sessions, facilitated cohort discussions, and spent three days listening. Here’s what we heard. 

Why AIIM hits different every year 

Most conferences give you keynotes and booths. AIIM gives you cohorts. Small groups of practitioners, together across all three days, working through real problems with no audience. Just an honest conversation about what's actually hard. 

That format produces something rare: an unfiltered signal. And what came through this year, cohort after cohort, was a striking degree of consistency. These practitioners from government, financial services, energy, airlines, and nonprofits were telling the same story. And it wasn't the one leadership wanted to hear. 

The people who can fix AI are already on your payroll 

The most consistent theme across every cohort discussion wasn't technology. It was whether the right people were in the right conversations. 

Information managers understand, deeply and practically, what it takes to make AI trustworthy at scale. They know which content is authoritative. They know where the governance gaps are. They know why the pilot that looked great in a demo will fall apart in production. Classification, metadata, lifecycle management, records policy — these aren't new skills being recruited in from outside. They're the skills that have been sitting inside your information management team for years, ready to be applied to exactly this problem. 

The opportunity is organizational. The practitioners who are making the biggest impact on their AI initiatives aren't the ones with the most technical knowledge. They're the ones who have successfully repositioned that knowledge as business-critical. They've made themselves indispensable to AI governance committees. They’re connecting governance to outcomes rather than risk. They're leading with "here's how we do this safely" rather than waiting to be asked. 

"A 1-year AI adoption strategy is a joke" wasn't a fringe opinion at this event. It was the consensus. Executive-set AI timelines are consistently disconnected from the data estate reality underneath them. The information managers who are changing that dynamic aren't pushing back. They're stepping forward, with a concrete plan and a seat at the table to match. 

When the tool works fine, but the information doesn't

Technology wasn't the dominant topic in cohort discussions. Change management was. Training, adoption, AI literacy, the skills gap: these were the pain points that came up consistently and urgently. 

One practitioner described the core problem perfectly. Their chatbot was pointed at a system of record that hadn't been updated. Users got exactly the wrong answer, delivered with complete conviction. The tool worked fine. The information didn't. 

This is the reality beneath most enterprise AI initiatives: content sprawling across file shares, email, M365, and legacy systems. Metadata missing or inconsistent. Retention policies that exist on paper but aren't enforced. Integrations with SAP, Salesforce, and other business systems that were never designed with AI in mind. Nobody is managing the data estate until something goes wrong. 

AI needs governed content. Not just available content 

A common refrain at the event: "Enterprise AI doesn't have a model problem. It has a content trust problem." True. But in the enterprise, that's only half the answer. AI is only useful when that content is permission-aware, lifecycle-managed, auditable, and connected to the systems where work happens. 

That's the difference between content that's accessible and content that's trusted. Between an AI that can retrieve a document and an AI that can act on it with confidence. 

This was the focus of our workshop, Designing Your AI-Ready Enterprise. Partnering with Deep Analysis analyst Alan Pelz-Sharpe, Tracy Caughell took practitioners through a hands-on blueprint for moving from AI strategy to executable planning. The room wasn't short on ambition. It was short on confidence that the organizational conditions for success were actually in place. 

At OpenText, we call this the governed source of truth for AI. It's what Content Cloud is built to create, connecting content across SAP, Salesforce, Microsoft 365, Guidewire, and the rest of your business applications, with governance built in, not bolted on. Lifecycle controls, retention, audit trails, and access management embedded in the workflow rather than layered on top of it. 

A modern, cloud content platform isn’t optional

Tom Grucza's breakout session, Your Content Isn't the Problem. Your Platform Is., cut to the heart of something practitioners are living every day. Most organizations are trying to run enterprise AI on infrastructure built to store documents, not power decisions. Legacy document repositories lack the context, the governance, and the business integration that AI needs to operate reliably at scale. 

The path forward is modernization: not just moving to the cloud, but connecting content into the ERP and CRM processes where work actually happens, embedding governance from day one rather than retrofitting it later. AI performs best when grounded in governed, connected content, not files in folders. For organizations working within the Microsoft ecosystem, that means bringing trusted, contextual business information directly into Copilot experiences, turning content chaos into clarity rather than just moving it to a new location. 

The practitioners at AIIM 2026 weren't behind. They were ahead of the problem.

Classification, records governance, metadata discipline. That's AI work. It always was. 

Every conversation we had in Baltimore this week confirmed that it's the work that makes everything else possible. The question is whether your organization knows that yet. 

The post What 300 information managers want you to know about AI appeared first on OpenText Blogs.