AI Transformation

I’ve witnessed again and again companies trying to implement a RAG solution only to fail in the process. Most of these projects don’t fail because the technology is immature. They fail because the documents fed into them are contradictory, outdated, or simply shouldn’t be there. The oversimplified flow of a RAG (Retrieval-Augmented Generation) solution looks like this: a user asks a question, the system retrieves relevant passages from a vector database, and those passages plus the question are sent to an LLM to formulate an answer. Why such initiatives most often fail is quite obvious to the external eye but not so obvious to the people leading them from within. ...

In my day-to-day work I see firsthand what companies are planning as part of their AI transformation. Some projects are genuinely exciting and could reasonably be called moonshots. Others, in my view, rest on a basic misconception about what an AI agent actually is. What I keep finding is that many companies are not really looking for an AI agent. They are looking for a business rules engine. When you ask them to describe the specific use case, they describe a deterministic process: a fixed set of steps and rules that must be followed the same way every time. Think of an invoice under a certain amount that needs three matching approvals and an approved vendor before it can be paid. That is classic rules-engine territory—predictable, auditable, and repeatable. ...