This summer I noticed that most of what I understood about AI had come from building with it, not reading about it.
So I started treating my own work as a set of live experiments.
I began by rebuilding my own knowledge system at the beginning of the year. For years I’d kept notes neatly sorted by topic. They were stored, but they weren’t really helping me think or move. So I turned the whole thing into a pipeline: what I capture, what I’m actively working on, what has shipped, and what needs to be distilled into something reusable.
At the start of summer, I built an Integrated Executive Virtual Assistant on top of it. It has a mission control deck that shows me, at a glance, what’s still untriaged, which projects have gone quiet, and what I shipped but never turned into learning. It also drafts in my voice, and it gets better from every edit I make.
Building it taught me more about directing agents than a year of reading about agents ever did. A few people I trust are testing it now. I’m still deciding whether to launch it.
During July, I also started shaping an AI consultancy and agency around training an AI-native workforce, business workflow automation, and applied AI systems. That led to conversations about embedding AI into the core of a business by transforming and analyzing data lost between human-to-human interactions, with concrete offers such as bulk document processing for enterprise workflows using local LLM infrastructure.
Finally, during August, I’ll be building a small booking and personal training system for my brother’s smart training room in Colombia. It starts as a website with a booking flow, then becomes a local prototype I can use myself as a personal trainer. He’ll launch it as a way to collect leads and convert people to a monthly plan with local trainers.
None of these projects is huge on its own. That’s the point. They’re small enough to build, specific enough to break, and real enough to teach me something when they do.
Also in August: What Building Teaches That Reading Cannot · The Sandbox Is How You Learn · The Skill Under the Skill
Daniel Arevalo is a Product Architect and strategic product design leader. He helps product teams turn AI capability into trusted, adoptable workflows, with a background in industrial AI, IT/OT cybersecurity, and high-stakes enterprise software for critical infrastructure.