Talking earlier with someone else who’s implementing agents, I remembered something most people don’t understand, and that was very clear to us: the difference between agents and assistants.
Assistants are when AI is helping a person work better. It’s a catalyst for the work — for better and for worse.
Agents are when it’s executing the work autonomously, without a person stepping in. In other words, an automated process. With AI, of course.
The difference isn’t subtle, and the name matters little. Feel free to call them agents or assistants. However, assistants can use more modern things like MCP, Skills, etc. There will be someone in the process validating what actually works. IF that person isn’t just acting as a proxy for the AI (copying and pasting what the model says, or “vibing”, cough cough).
If it’s for more autonomous agents, I’d go as DETERMINISTIC as possible. That is, classic automation. I’d only put AI where it’s really needed. In most cases: understanding unstructured information. For everything else, don’t let the model decide — if you want something that always runs the same way.
At least for now…
Originally written in Portuguese and published on LinkedIn on October 21, 2025. I translated it into English with the help of AI (Claude): it let me bring my LinkedIn posts to people who don’t read Portuguese without rewriting each one from scratch. I reviewed the translation to keep my original meaning and tone.