About Tom Meli
In March 2016, I stayed up to watch DeepMind's AlphaGo play Lee Sedol. I had studied Go for years, so the match made AI's capabilities concrete for me. I wanted to understand what this technology could let one person do and help other people use it well.
Today, I coach capable professionals who want AI to carry more of their work without losing the context or standards that make the work theirs.
AI should help you do your work better. My goal is for you to leave with a useful routine you understand and enough practice to adjust it as your needs change.
Background
Before AI, I spent years learning where people get stuck with difficult skills and which next step gets them moving again.
I worked as a wilderness instructor, then coached students through high-pressure test prep and college essays. The work taught me to find the exact place someone is stuck and choose the next useful step.
I built a teaching company and an online learning platform for my students. The challenge was to preserve the parts of one-on-one teaching that helped each student make progress.
I watched Lee Sedol lose to DeepMind's AlphaGo at three in the morning. I had studied Go for years, so I knew what I was seeing. That night changed what I wanted to work on.
I work one-on-one with people who have tried AI but have not found a dependable place for it in their routine. We choose one useful task, test the routine on their own work, and save the instructions or checks that improve later runs.
How I work
We choose tools only after we understand what matters to you, where your information lives, and where the work currently breaks down. The setup should fit the way you already think and work.
We choose checks based on what can go wrong. A calculation may need a direct test, and a sensitive action may need approval. Lower-risk drafts can run with lighter supervision. When a result misses, we save the useful correction so later runs start with a sharper standard.
I teach as we build. By the time we finish the first task, you understand enough to adjust it, improve it, and start the next one on your own.
You know which information the assistant can use, what it may change, and which consequential actions require approval. Privacy, sources, and responsibility are clear from the start.
What you will find here
Practical guides for connecting your tools, setting up memory, and designing reusable prompts you can adapt and build on.
A daily AI term with plain definitions, common confusions, and quick practice when you want to check your understanding.
We find the first task where AI clearly helps your work, build it until it is reliable, and go from there.
