Enterprise Knowledge as the Strategic Foundation for AI

Fluid Topics is the undisputed leader when it comes to reinventing how end-users, field technicians or support agents can access and use technical documentation. And as you imagine, AI plays a big role in this. Because here’s the catch: enterprise AI isn’t hitting a technical wall, it is hitting a knowledge wall. Across organizations, systems are being deployed faster than the information they rely on can keep up. Information remains scattered across multiple sources often inconsistent, poorly structured, and difficult for machines to interpret. As a result, AI struggles to deliver reliable outputs, not because it lacks intelligence, but because it lacks the right foundations.

Fluid Topics operates at the center of this challenge. Leveraging more than two decades of expertise in semantic search, artificial intelligence, and content accessibility, the company has built a Product Knowledge Platform that transforms fragmented information into usable, contextualized knowledge for both humans and AI systems. Fabrice Lacroix, Founder and CEO, shares his perspective on how organizations must rethink knowledge in the age of AI.

Q. Fluid Topics works with companies to better manage and use their knowledge. How would you explain, in simple terms, what you bring to your clients today?

A. Fluid Topics is a SaaS solution known as a Product Knowledge Platform. It is designed to help companies turn fragmented information into a unified knowledge center, a single source of truth. Our platform consolidates content from multiple systems, enriches it with context, and makes it available across various touchpoints such as documentation portals, chatbots, and support tools. The objective is straightforward: whenever a user or an AI agent needs product knowledge, they can instantly access the right information, in the right context, every time.

Q. Many organizations generate a huge amount of content, but struggle to actually use it. Where do you see the biggest gaps today?

A. The biggest gap today is content fragmentation: across tools, formats, and teams, with inconsistent structure and governance. This makes it difficult not only to access information, but to trust which is correct.

There is also a growing mismatch between how content is created and how it is consumed. Documentation is still primarily written for humans, while it is increasingly being used by AI systems.  AI requires a different granularity and a richer content to provide the context and fundamentals that humans would naturally include.

Without that, even high-quality content becomes difficult for machines to interpret and use reliably. As a result, companies are sitting on valuable knowledge that is technically available but operationally unusable, especially for automation and AI-driven workflows. That gap between availability and usability is where most of the missed opportunities lie.

Q. You often say that knowledge is becoming a strategic asset. Why is it now a key topic for business leaders?

A. Knowledge is strategic because it directly drives business results. In a world where AI can automate decisions and actions, the success of those processes depends entirely on the accuracy and accessibility of the underlying information. Outdated or incomplete content can lead to errors at scale. On the other hand, well-governed knowledge becomes a competitive advantage as it enables faster decisions, better customer experiences, and more efficient operations. That’s why business leaders are paying attention now.

Q. With the rise of AI, companies are rethinking how they use their internal information. What is changing in the way organizations approach knowledge?

A. What’s fundamentally changing is the role of knowledge itself.  It’s no longer just there to inform people. It’s now expected to power AI systems that can assist customers, solve issues, or suggest the next best step.

This shift is pushing organizations to move beyond traditional approaches like document storage or basic search. Instead of asking, “How do we give employees access to information?”, they’re asking, “How do we make this knowledge usable, reliable, and actionable for both our employees and AI agents?” That requires much stronger structure, clear ownership, and well-defined sources of truth.

In short, organizations are shifting from managing content to engineering knowledge, designing it not just to be read by humans, but to be directly consumed and acted upon by intelligent systems in real time.

Q. A lot of companies have tried to centralize their data, sometimes with limited success. What are the common mistakes you see?

A. I think many companies start with the right intention, creating a single place for knowledge, but they underestimate the complexity behind it. What usually happens is that content gets centralized without a truly coherent structure behind it. People and now AI systems struggle to understand what’s authoritative.

Then there’s the issue of keeping content up to date. In many source systems, data is changing every minute or second, and information becomes immediately outdated once exported. At that point, you’re no longer looking at a source of truth, but at a snapshot that’s already obsolete.

And a third challenge that’s often underestimated is security. In many cases, when content is extracted and moved, the original permission models don’t come with it. That introduces risk and potentially allows unauthorized users to access sensitive material.

So, in the end, the difficulty isn’t centralizing content, it’s maintaining something that is structured, up-to-date, and secure. Without those three elements, centralization can actually add another layer of complexity instead of solving the problem.

Q. If you had to give one piece of advice to companies looking to turn their knowledge into a real performance driver, what would it be?

A. Focus on the content, not on technology. Don’t build it; buy the right tools. It’s changing too rapidly anyway. That’s the vendors’ job to bring you the latest technology. Put your energy into identifying the users you want to augment, the processes you want to automate, and figure out the knowledge that’s needed for that. The information that’s missing so that AI can make the right decisions.