Philip Brittan is CEO of Bloomfire, pioneering Enterprise Intelligence solutions for Fortune 500 companies.
getty
A few months ago, we ran our own internal knowledge base through a health-scoring tool we’ve been developing at my company. It was the kind of structural audit we help enterprise customers think about every day. Our content library scored 91 out of 100. By any measure, that was a strong result. And yet the analysis surfaced 31 concepts our documents failed to ever define. One of those concepts? Our core AI product. It was referenced in 10 documents, but not one explained it.
I’ll be honest: That finding stung. We build the accountability layer. We think about knowledge governance for a living. We still had this problem, invisibly, until we looked.
If that can happen at a knowledge management company, it can happen anywhere. Almost certainly, it’s happening at your organization right now. The question isn’t whether your knowledge foundation has gaps. It’s whether you’re about to pour another round of AI investment on top of them.
Before you do, I’d suggest three questions. Drawing on the themes and accountability framing that run through my company’s broader “Guide to Enterprise Intelligence Systems” report—commissioned by my company and produced by Dr. Anthony Rhem—here are three questions I’d ask. I find them useful not because they’re comfortable but because they’re clarifying.
Who bears the cost when it breaks?
This question hits first because most organizations haven’t named the number yet. Our “Value of Enterprise Intelligence” report, which surveyed more than 10,000 employees across 115 companies, puts the cost of knowledge management at roughly 25% of annual revenue—approximately $2.4 billion for a mid-size manufacturing company in an example our report included.
Trace where that figure actually lives: in the hours employees spend hunting for information that should surface immediately, in the decisions made on outdated guidance, in the onboarding that takes twice as long as it should because institutional knowledge never got captured. I’ve seen this pattern repeat across industries. The cost doesn’t announce itself. It distributes across teams, workflows and outcomes in ways that compound slowly enough to miss until they’ve grown too expensive to ignore.
What AI does is accelerate that distribution. A knowledge governance failure that once produced a slow leak becomes, at AI speed and scale, something much harder to contain. Garbage in, garbage out remains an inviolable principle. And at AI speed, the garbage moves faster than any team can chase it.
The answer isn’t slower AI adoption. It’s investment in the foundation before stacking capability on top of it.
Who holds accountability for the outcomes?
Early in my career, I spent years at global organizations. In financial data, the accountability question never stays abstract. Stale data carries a name and a cost: a wrong price, a missed trade, a regulatory violation. The chain of responsibility—who maintained the data, who certified it, who acted on it—wasn’t bureaucracy. It was the architecture of trust that made the entire system function.
Most enterprises have never applied that same discipline to their knowledge assets. Knowledge grows, ages and fragments without anyone owning the life cycle. When AI surfaces something wrong with confidence, the question “Who holds accountability for this?” often has no answer because no one was ever asked to answer it.
Ask yourself: If an AI-generated response shaped a customer interaction, pricing decision or compliance filing at your organization last week, could you trace it? Could you name who owned the knowledge it drew from, when that knowledge was last validated and who would know if it had gone stale?
In my experience, most organizations can’t. That’s not a technology gap. It’s an organizational design gap, and it has organizational design solutions: Establish knowledge ownership roles, content life cycle policies, review workflows and governance structures that treat organizational knowledge with the same rigor applied to a data feed.
Can you explain where your AI’s decisions come from?
Most enterprise AI retrieves whatever it finds in the underlying knowledge base, assembles it into context and responds with the confidence of something that knows exactly what it’s talking about. Whether that knowledge is current, accurate or internally consistent is not the model’s problem. It becomes yours.
This is why AI disappointment numbers stay so consistent across surveys. McKinsey found that 60% of organizations said knowledge and training gaps, not technical limitations, are the primary barrier to responsible AI implementation. AI budgets reportedly rose 38% last year. Forty percent of that spend underperforms.
The organizations asking, “Why isn’t our AI working?” tend to examine the model. But the answer might live in the knowledge base the model draws from and in whether anyone can trace where a given output came from, why it surfaced and whether the source deserves trust. Transparency isn’t a feature. It’s the mechanism by which you catch the problem before it scales.
Where To Start
The three questions above point toward a diagnostic, not an overhaul. Here’s how I’d approach it:
1. Run a knowledge audit. Map where critical knowledge lives, who owns it and when it was last validated. The gaps tend to surface quickly (ours did).
2. Assign ownership. Every knowledge domain needs a named owner with accountability for accuracy and currency. Without that, governance stays theoretical.
3. Measure what you can’t see. Track how often employees bypass AI tools, duplicate work or make decisions on outdated information. Those metrics make the investment case for fixing the foundation.
None of this stays theoretical. Those 31 undefined concepts in our own knowledge base were a governance failure at a company that exists to solve governance failures. The question every enterprise leader should sit with isn’t whether to invest in AI. It’s whether the foundation underneath it deserves a decision staked on it.
Most of the time, you won’t know until you look.
Forbes Business Council is the foremost growth and networking organization for business owners and leaders. Do I qualify?
