Why demos survive and deployments struggle
What changes when an agent meets real customers, messy records and an ordinary Monday.
Read articleBlog
Practical reading for the people building, buying and running AI agents.
Articles in English
When the customer changes the plan
How to handle corrections, implied needs and requests that evolve mid-conversation.
Read article
A closed chat is not a solved problem
Measure the customer’s outcome, including everything that happens after the chat.
Read article
Your data needs more than a search box
Help agents understand records, exceptions and the context your team takes for granted.
Read article
The API failed. The agent said “done.”
Build for timeouts, partial actions and the gap between a reply and a result.
Read article
Tests that change what you ship
Choose cases and measures that help your team make a real release decision.
Read article
Give agents the access the job needs
Keep useful autonomy without handing over every permission a human happens to have.
Read article
The price of actually finishing the job
Count retries, review time and cleanup when you measure what an agent costs.
Read article
Who owns the agent after launch?
Keep responsibility, feedback and improvement connected after the project handover.
Read article