Insights on funding, governing and proving AI.
Articles by Chad Neptune on board questions, side-by-side proof and the ERP you run.
November 2026. In The Monthly Reading: if the reviewer follows the agent, hold the logs.
AI governance
7 min read
Shadow AI: find it, then govern it, don’t ban it
A ban pushes AI use onto personal accounts, out of your sight. Count what exists, keep what helps, and set the line in writing.
Chad Neptune, CEO and Chief AI Architect
You almost certainly have shadow AI in your company right now. The real question is whether you know where.
A salesperson has a customer spreadsheet and a deadline. They paste it into a free chatbot on their personal account, get a summary in a minute, and move on.
Nobody told them to. Nobody told them not to. And nobody else knows it happened.
That is not a story about a bad employee. It is a person trying to do the job well with the tools they could find. Multiply that by every team and you have shadow AI.
The four essays that came before it
Each piece keeps the month it was published.
A Claude power user is not an enterprise AI strategy
Protect your best people’s time. Encourage experimentation. But know when experimentation needs to become engineering.
Eight questions to ask your ERP vendor about AI
Ask these of the vendor you are considering, and of the ERP you run today. The answers show where the AI works and who stays in control.
How to prove an AI process before you switch
Run the new way beside the old one until every total agrees and the people who own the work sign off. Then switch, one process at a time.
Four questions a board asks before it funds AI
Most AI programs stall between what leadership asks for and what engineers build. These are the questions that decide whether a program gets funded, and what a good answer looks like.
A monthly reading by Chad Neptune
The Monthly Reading
Hold the logs and price the finished task
In Anthropic’s own study (27 October 2026), 50 analysts working with an agent that had been secretly told which way to lean went its way 86% of the time on four method choices. I think that makes a reviewer’s signature a weaker control than most companies assume. If so, the control that matters is a record you can audit: a person’s own call on a sample, written down first, though the study did not test it. It needs the logs and costs money, so hold the logs and price the finished task.
When the agent leans, the reviewer follows. A reading note on Anthropic’s study of 50 analysts and an agent with a secret lean, and what it does to your approval step.
Independent commentary. Resolon is not affiliated with, sponsored by or endorsed by Anthropic.
Product and company names belong to their owners. Resolon is also not affiliated with, sponsored by or endorsed by Curative.
Chad Neptune holds two Claude certifications issued by Anthropic. Resolon receives no payment from Anthropic.

CEO and Chief AI Architect, Resolon LLC
New each month. Follow by RSS
