Insights
Notes on making AI work in a real company.
Why AI seats go unused, why pilots stall, what the insurance questionnaire asks, and how TNT runs itself on its own stack and agents. Written by Thiago Damasceno, TNT's founder.
Posts
Read the latest.
Newest first. Each post carries one of the four topics below.
-
The AI you already pay for
Why Copilot seats go unused, and what to fix first
Copilot answers from what each person can open. When that's too little or too much, people stop using it. Fix access, then permissions, then one task.
-
The pilot that stalled
Why AI pilots stall: the four stages, and how to find yours
A pilot often stalls because it assumed a stage the company hadn't reached. The four stages, Pre-AI to Native, and the signs that show where you are.
-
Behind the build
How our intake works: four ways in, one record
The website form, info@, Bookings and a Microsoft Form all land on one Dataverse record, acknowledged in seconds. What's live and what's still in build.
Topics
Four topics, all about making AI work.
Three start from a problem you may already feel. The fourth shows how TNT runs itself on the same stack.
The pilot that stalled
Why AI projects stop after the demo, and what has to be true about data, ownership and the workflow before the next one.
The AI you already pay for
What your licenses already include, why AI seats go unopened, and what to fix first.
Security and compliance for the AI era
The insurance questionnaire, the policies and the evidence a company needs before agents touch its data.
Behind the build
How TNT runs itself on the Microsoft stack it builds on. What's live, what's in build, and what broke on the way.