Source code is structured and easy to validate. How can you validate meeting notes and test commercial analysis? During my recent workshop, I found that Project Managers were more interested in AI agents than the Developers in the room. The reason became obvious: with a little help, PMs can improve their effectiveness in several distinct areas. The problem Software development is relatively easy to automate, but project management is more versatile and requires a different approach.

The local dim sensor failed to provide proper readings; there was no way to calibrate it or fix it otherwise. Since that was a stationary installation, I used the following workaround: 1️⃣ For those specific GPS coordinates, I created a lookup file containing sunrise and sunset timestamps for every day of the year. 2️⃣ I used UTC timestamps to avoid issues with daylight saving time updates. 3️⃣ When an event was detected, a simple script calculated the day/night-time value and enriched the AI agent’s input.

In order to scale my single-person enterprise, I had to create an agentic development department. I’ve been working on this idea for a few months, overcoming various ups and downs (including software and hardware limitations). This setup was intended as my internal tooling to support daily development activities. My goal was to automate the commercial-grade software development, testing, and auditing process. I still manually review the final code changes, but I am not involved in iterative development.

The Problem Statement It rarely happens that an AI agent decides to drop the database to “recover space”. But if your agent gets too creative, how can you prevent it from doing harm? To truly leverage the potential of AI agents, we need to grant them significant autonomy. I’ve been experimenting with various approaches, and my agents went the “wrong path” more than once. Typically, those incidents were harmless (my test agents run on a dedicated, isolated server).

The Problem Statement I wanted my agent to help me manage my emails, but I do not want it to be proactive and send emails in my name (which it did during my initial research phase). AI tried to spam potential customers - not what I wanted!!! Unfortunately, the default Gmail API is not very selective - the modification access also allows sending emails. The Implementation GPT5.5 suggested building a custom MCP extension to add extra guardrails.

I am impressed by how that model can navigate web pages and analyze data. Unfortunately, when that model generates a report, it makes even more spelling mistakes than I do. On the other hand, Bielik-11B-v3.0-Instruct is not as smart and capable as the Holo model - there is no way I would use it as the main “brain” of my local agent. But guess what? There is a way to make those two models cooperate in a way that each acts in the area it absolutely dominates!