If you’ve run an AI project in the past year and felt let down by the results, the problem is rarely the tool. It’s whether your content is accurate, connected, and ready for AI to reason over. That’s what AI content management solves.
That’s the focus of a recent panel discussion from OpenText and Microsoft, “From chaos to clarity: Building the data foundation for AI with Content Aviator and Microsoft Copilot.” Tom Grucza, Senior Product Marketing Manager at OpenText, and Thomas Matthew from Microsoft’s Enterprise Partner Solutions team break down why intelligent content management matters before you scale tools like Microsoft Copilot, and how to build that foundation. Here’s what they cover, and what you’ll take away when you watch.
Why AI initiatives stall
The barrier to enterprise AI value sits below the model, in the data foundation. Industry surveys on enterprise AI readiness cited in the webinar point to three recurring problems:
- 52 percent of organizations struggle to access data that’s complete, accurate, and ready for AI. If the data isn’t reliable, the output can’t be either.[1]
- 42 percent of leaders report low confidence in AI results because information is fragmented, siloed, or disconnected from other systems. AI can’t reason well when it sees only part of the picture.[1]
- 89 percent cite concerns about protecting sensitive, regulated information when using AI.[2]
Microsoft’s own analysis, shared at Ignite, points to the same root cause: content quality and readiness across structured, semi-structured, and unstructured sources have to be addressed before that content feeds AI and agentic systems.
Enterprise data is proprietary, regulated, and business critical. It isn’t just the structured records you picture in a standard AI use case. It’s the contracts, invoices, HR records, engineering documentation, and operational content that run the business. Some teams try to work around the problem by copying everything into AI lakes or models, but that often raises risk. It breaks lineage, creates compliance exposure, and introduces data freshness issues. The better approach makes content AI-ready where it already lives, so it stays governed, connected, and secure.
The dark data problem
Even with strong information management in place, much of your most valuable content stays dark. Documents sit in legacy repositories, file shares, archives, and departmental systems. The information isn’t bad. It just isn’t AI-ready. It lacks consistent metadata, clear ownership, and the business context that an AI assistant needs to deliver deeper insight.
What makes content dark for AI is what makes it dark for people. Can you find it? Do you even know it exists? Most organizations have a handle on perhaps 10 percent of their content and assume that’s the whole picture[3]. In reality, a document locked inside Salesforce or SAP that no one can reach is a decision made with half the facts, whether the decision-maker is a person or an AI assistant. As a result, only a portion of enterprise content can be trusted for AI today. The rest stays stored but inaccessible, without context, and outside AI governance.
What AI content management looks like in practice
A modern foundation does far more than store files. Enterprise content management built for AI creates a single source of truth, one place where content lives so that any approved person or AI assistant has the full context to make the best decision.
That foundation works because of what runs behind the scenes:
- Metadata that describes the content and keeps it trusted and governed.
- Knowledge graphs and ontologies that relate a customer record to its contracts, purchase orders, packing slips, and paid invoices in one complete view.
- Governance and lineage are treated as first-class, so retention, security, and access controls travel with the content.
These capabilities have lived in intelligent content management systems for years. What’s new is that AI assistants can finally tap the relationships already stored there and turn them into insight. The work people once did by hand, gathering related documents to understand a case, now happens in context and at speed.
This is where content integrations make the difference, and where OpenText and Microsoft complement each other. OpenText intelligent content management acts as the single source of truth for unstructured business content. Microsoft brings a logical control plane that spans SharePoint, Teams, OneDrive, and structured sources in Microsoft Fabric and OneLake, with identity, sensitivity labels, and retention policies layered on top. Together, they give AI a connected, policy-aware foundation to reason over.
Content Aviator and the Microsoft 365 integration
OpenText Content Aviator is an AI assistant that analyzes and summarizes the unstructured content managed in an OpenText content platform. Ask it to tell you everything about a customer before a call, and it pulls from dozens of documents, plus connected sources like Salesforce or SAP, and brings back a synthesized answer.
The news from the webinar: OpenText recently released an AI-to-AI integration between Content Aviator and Microsoft Copilot. This Microsoft integration lets people who live in Copilot invoke Content Aviator wherever they work, including a dedicated Copilot interface or a Microsoft 365 app like Word, PowerPoint, or Excel. Copilot calls Content Aviator to search and analyze content governed in the OpenText platform, then delivers those insights right inside the Microsoft 365 experience.
The principle is simple. If OpenText is the steward of your single source of truth, any AI assistant that needs that content should be able to reach it, safely and in context. The integration with Microsoft 365 is the first of these connections.
How to get started
You don’t modernize everything at once. The panel laid out a practical path from siloed systems to trusted enterprise AI.
- Start small and discover. Pick one high-value, content-driven business process. Map who’s involved and what decisions get made, then align your AI goals with real outcomes and define what success and ROI look like.
- Validate with real data. Prove the use case with a 30-day trial using your own content, so you can see how a governed single source of truth, Content Aviator, and Copilot perform together before you commit.
- Scale what works. Once you’ve proven the combination, find the repeatable pieces and extend them across the enterprise.
Exception-heavy workflows make strong first candidates. Where a share of transactions needs human review, you can automate the standard steps and bring Content Aviator or Copilot in to help the person in the loop decide faster.
There’s a cost angle worth raising with your account teams as well. Joint OpenText and Microsoft customers can often apply their Microsoft Azure Consumption Commitment (MACC) to move an existing content management system to Azure, which makes modernizing the foundation simpler and more affordable.
Build the foundation, then scale the AI
As the MIT research referenced in the webinar puts it, high achievers align their data and AI strategies tightly with business outcomes. The lesson is to keep the business process, the outcomes, and the people at the center, rather than turning on technology for its own sake. AI content management, a true single source of truth, and AI assistants working across it are what turn content you already own into trusted answers.
Watch the full webinar to hear Tom and Thomas walk through AI content management in detail. Ready to see it on your own content? Start a 30-day Content Aviator trial, upload up to 100 of your own documents, and ask your own questions.
Sources:
[1] Foundry Research MarketPulse Survey, “The Role of GenAI in Modernizing Content Management, May 2025
[2] Forrester report, “Results Of Forrester’s Automation Survey, 2024
