Teaching the Machines to Teach Us Back
Introducing My AI-powered API Coaching Project
A new kind of learning companion
For years I’ve been writing and teaching about APIs, hypermedia, and the architecture of complex systems. But lately, I’ve been working on something different, something that sits between my books, my talks, and the code itself
.
I call it AI Coaching and I’m assembling a growing collection of interactive “coaches “ that guide you through the thinking, modeling, and building of great systems. Not code generators. Not chatbots. But mentors in miniature.
Each stand alone coach focuses on a specific domain. Here are some examples I have created so far:
API Story Coach – helps you craft an API narrative before you ever open an editor.
Vocabulary Coach – aligns your terms, language, and domain meaning.
Security Coach – reviews your design against good governance patterns.
Persona Coach – helps you design with empathy, not just requirements.
Engelbart Coach – explores how technology can augment human intellect rather than replace it.
...and more to come – as new contexts, challenges, and opportunities emerge.
All of these coaches live in a shared framework I’ve been building, a library of Context Kits that make it easy to create your own specialized coach for any project, team, or organization. Each kit defines ethics and safety boundaries, step-by-step guidance, and the expected artifacts (from Markdown notes to OpenAPI specs to Node.js prototypes).
Why build AI coaches at all?
Using AI as a coach or mentor isn’t new. There are already several companies exploring this idea, from general-purpose productivity mentors to executive training bots. But most of those efforts paint in broad strokes. They focus on business operations and culture.
My project takes a different path. I’m focusing on the API design space, where AI can serve as an informed collaborator for developers, architects, and teams working to build resilient, evolvable systems.
The aim isn’t to automate and eliminate work but to expand and improve creative problem-solving skills and to strengthen the human empathy that shapes their designs.
The balance that matters
One of the driving ideas behind this project is maintaining the strength of two worlds: the human side (e.g. judgment, interpretation, and choice) and the tool-driven side (consistency, precision, and automation).
Right now I see too many genAI tools blur the line between human creativity and machine efficiency. I’d rather strengthen the bridge between them.
These AI coaches I am implementing sit right along that human-machine bridge. The coaches don’t make design decisions, but they keep humans and machines in dialogue. They suggest, lead, and inform API builders. More importantly, they remind us to when trust our intuition and when to lean on structured assistance.
In short:
The aim is not to replace your skills but, instead, to amplify them.
What’s next
Over the coming months, I’ll release these coaches (and their README files) at my github repo, one by one. Initially they will all be free. At some point, as the collection grows and matures, I’ll be releasing a book of them (along with other materials). The plan is for this book to also include guidance on how you can apply this AI Coaching Context model to your own organizations. And maybe even publish and share your coaches with the wider community.
If this project sounds intriguing, I hope you’ll join me in trying them out, providing feedback and experimenting with your own Generative AI Coaching projects.