Ed is CEO of Intelligo. A former prosecutor, he leads a risk intelligence platform for due diligence, combining AI and expert analysis.
I have spent my career on two sides of the same problem. Building businesses that ran on data, and before that, prosecuting cases that lived or died on whether a fact could survive being torn apart on cross examination. Both taught the same lesson: the volume of information you hold is not the asset. The asset is what you can connect and what you can defend.
Most firms have this backward. And they have it backward three times over.
Walk into any serious organization and you will find multiple intelligence functions running in parallel, under different names, in different systems, owned by different people.
Business intelligence turns the firm’s own data into decisions. Revenue, cost, operations, performance. Competitive intelligence turns outside data into decisions. Competitors, customers, markets, where the field is moving. Risk intelligence turns the data of threat into decisions. Counterparties, exposures, sanctions, litigation—the things that can go wrong.
Multiple names. Multiple stacks. Multiple teams. One motion. Every one of them takes data the firm can already access and works to turn it into a decision. They are not distinct disciplines. They are one discipline, practiced in rooms that do not talk to each other.
The counterparty your deal team is evaluating is the same entity your competitive analysts profiled as a market mover and your business intelligence shows you already pay as a vendor. One entity. Multiple pictures. Multiple processes. Nobody holds all views at once, so nobody sees the thing that only appears when you do.
The unlock is not a better version of any one of those functions. It is the walls coming down.
The Knowledge Gap
Frank Knight, the University of Chicago’s legendary economist whose 1921 book Risk, Uncertainty, and Profit became a foundational text in economic theory—and whose framework continues to shape how risk managers and regulators think about uncertainty—drew a distinction that has held for a century: risk is what you can measure and price; uncertainty is what you cannot. Knight’s “line” was about the limits of probability itself. Inside a firm, the same split shows up in a more fixable form: not unmeasurable uncertainty, but measurable risk that nobody bothered to measure because it was scattered across rooms that don’t talk. Most organizations sort their intelligence by department, credit, compliance and market research. Knight’s framing is more useful: knowable vs. unknowable. The knowable risk that goes unpriced is almost always hiding in the gap between functions. It exists in the firm’s data. It was never connected, so it was never priced.
The walls between those functions were not a conceptual error. They were a practical response to volume. Unifying growing oceans of data into one picture, keeping it current, at the scale data now arrives, that was more analytical labor than any organization could carry by hand. So firms split the work into rooms and accepted that the rooms would not talk. The idea of one intelligence was sound. The means were missing.
The means are no longer missing. The rooms remain.
Unified Intelligence
That is the real opportunity AI offers this field. Not faster reports in each room, but one intelligence across all of them. But here is where a prosecutor’s instinct becomes essential, because not all AI is built the same way.
Probabilistic systems, designed to predict the most likely output from patterns across vast data, are genuinely powerful. They are also, by design, unable to tell you exactly which document produced a specific conclusion. They approximate. In a deposition, approximation is not a defense. In a regulatory inquiry, “the algorithm suggested so” is not an answer.
Deterministic architectures work differently. They force every finding to map back to a specific, verifiable source—not a statistical inference, but a traceable fact. That is the distinction that makes unified intelligence defensible rather than just fast. The evidentiary foundation is not a feature. It’s the architecture.
The Standard
I know what happens when a conclusion cannot be traced. Not in the room where it was produced, but eighteen months later, under adversarial conditions, when someone with a strong incentive to find fault is asking questions. A connection you cannot trace to a document is not a finding. It is a guess that learned to sound like one.
So here is the thesis plainly. A firm does not have a business intelligence problem, a competitive intelligence problem and a risk intelligence problem. It has one intelligence problem, cut into multiple rooms by tooling that no longer constrains it. The data already exists. The concept already exists. What was missing was the means to hold all of it at once; traceably, at the volume the modern firm produces.
Intelligence is not a department. It is the connected sum of what a firm knows, made visible, kept current, traceable to source and pointed at a decision. The firms that understand this will stop running intelligence functions that cannot see each other. They will run one. The rest will keep filing what they know in separate rooms, and keep being surprised by things that were sitting in them all along.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?