Notes on enterprise intelligence.
Analysis of enterprise AI, institutional knowledge and the conditions under which agent systems reach production.
What an Air-Gapped Enterprise AI Deployment Involves
Air-gapped AI is asked about constantly and described precisely almost nowhere. Here is what running an enterprise intelligence layer with no outbound network route actually requires.
What Frontier Intelligence Means for Enterprise AI
Frontier models hold general intelligence. Frontier intelligence is institutional: the knowledge, judgement and procedure of one enterprise, working above the model. Here is what that means in practice and why it changes the economics.
Why Your Enterprise AI Bill Grows Every Month While Capability Stays Flat
Enterprise AI spend compounds. Enterprise AI capability usually does not. The reason is architectural, and it is the half of the problem most knowledge-layer arguments never address.
AI Agents for Regulatory Compliance: Why Context Is the Bottleneck
AI agents are being deployed across compliance functions at speed. The bottleneck is not model quality. It is the institutional context that makes the difference between a generic compliance tool and one that reflects how your organisation actually operates.
Why Basic RAG Fails Enterprise Workloads
Retrieval-augmented generation works well in demos. In production enterprise environments, basic RAG fails in predictable ways that more sophisticated document intelligence solves. Here is what those failure modes look like.
What Are AI Agent Skills? The Missing Layer in Enterprise AI
AI agents can answer questions and generate content. Skills are what let them execute work: following the specific procedures, decision rules, and methods your organisation actually uses. Here is what they are and why they matter.
How to Measure Enterprise AI ROI Before Scaling
Most enterprise AI ROI calculations are built on demo performance, not production reality. Here is a framework for measuring what AI delivers in production, including the costs most calculations ignore.
The Consulting Firm's Guide to AI That Compounds
Most consulting firms are using AI as a productivity tool. The firms pulling ahead are using it as a knowledge infrastructure. Here is what the difference looks like, and why it matters more every month.
Why Your Company Knowledge Base Is Failing Your AI Agents
A knowledge base built for human search does not work for AI agents. Here is what needs to change, and why most companies discover this too late.
AI Agents in Lending, Collections and Client Management
Lending, collections, relationship management and portfolio work all depend on knowledge that is never written down. Here is why AI agents deployed into those functions plateau, and what closing the gap requires.
How to Capture Institutional Knowledge Before It Walks Out the Door
Most companies lose institutional knowledge faster than they capture it. Here is a practical framework for extracting what experienced people know, and making it available to AI agents and the teams that follow them.
Why Your AI Agents Have No Memory, and What It Costs
Every time your AI agent starts a new session, it forgets everything it knew about your company, your clients, and your decisions. Here is why that is a structural problem, and what fixing it actually looks like.
What Enterprise AI Knowledge Management Requires
Enterprise AI knowledge management is not a search engine or a chatbot. It is the system that decides how much of your company's intelligence your AI agents can actually access and act on.
Why Professional Services Firms Need an Intelligence Layer
Professional services firms sell knowledge that lives in people. That makes them uniquely exposed when those people leave, and uniquely well placed to benefit from a layer that retains what the firm already knows.
Why 95% of Enterprise AI Pilots Fail to Scale
An MIT study reported in August 2025 found that 95% of enterprise generative AI pilots never scale beyond the initial experiment. The primary cause is not the technology. It is the missing layer between company data and the agents that draw on it.