AI products.

I design and build applied AI products — tools for human–AI collaboration, agent workflows, permissions and trustworthy operation.

What recurs across the work

The same questions come back in every AI product: who the user is and what they already know; what the tool may do on its own and what needs a human’s yes; what the user can see of what the model did; what happens when it fails; what evidence the tool leaves behind; and how you test any of that without a live model in the loop. OpenBots is the fullest answer so far. The smaller tools are earlier attempts at the same questions.

Lead proof: OpenBots

A native macOS app that turns Claude into a team of persistent, named teammates — durable local state, explicit approvals, Swift 6, MIT. Preview build: the executor for consequential actions is disabled.

OpenBots

OpenBots — Scout’s thread in a fresh workspace, with the teammate family in the sidebar

Also: AI, Applied.

A searchable public library of real AI use cases from the Wharton Alumni AI Studio community — convened and built by me.

Earlier tools

  1. Active Listening AI

    An AI that listens and helps you work through your thoughts without offering advice. No clock, free, available 24/7.

  2. Easy Local LLM Guide

    What local LLMs are, how they protect your privacy, and how to install the popular ones on a Mac.

  3. Privacy Protection (Meta-Data Eraser)

    Remove hidden information from your files before sharing them.

How the building gets done

I design the product and make the technical calls — architecture, stack, interactions — then direct AI coding tools that build to my spec. I review and test everything that ships.

Hire me for this

I build AI tools for your team. Working software, designed with the people who’ll use it. It takes the repetitive work off their plate; they keep the judgment.

Let’s work together

Or write to lorenzo.colombani@live.fr. Available on retainer, for a few clients at a time.