Four Lanes: Where Enterprise AI Actually Stands in September 2026
A field guide to context, sovereignty and the economics of enterprise AI.
Ideas for AI
that has to work.
From Forward Deployed Engineers to Forward Deployed Software.
Engineers can earn the first use case. Software should make the next one easier.
Read the article22 articles
Your AI Agent Is Slowly Turning Into a DAG
What happens when the reliable parts of an agent become conventional code.
Agentic development did not fix delivery. It exposed it
Faster code is only the beginning of trustworthy production delivery.
The Determinism Problem
Where agent reasoning ends and deterministic execution begins.
False Summits: Why Enterprise AI Keeps Looking Finished
Why enterprise AI keeps looking finished before it is ready for production.
The Work Is Inside the Parentheses
The next frontier is systems that know what they are checking for.
What Woke the Agent?
Use tokens to manage business events, not to check for silence.
From Forward Deployed Engineers to Forward Deployed Software
Engineers can earn the first use case. Software should make the next one easier.
Stop Hiring Your Agents
The enterprise needs a new approach to work, not digital employees.
The Single-Provider Trap Is Coming to Enterprise AI
The architectural choices behind enterprise AI independence.
Evaluation Is How Agentic AI Earns the Right to Act. Here's What We're Finding.
How evaluation establishes the boundaries of agent autonomy.
Memory Is the Next Frontier for AI. Here’s What We’re Finding.
What persistent memory means for agents working inside an enterprise.