Why this site exists
Building software used to mean knowing how to code, or having the money to pay someone who did. Building with AI has changed that. It is a real, learnable skill, and with a clear problem and some persistence, you can build the tool you need yourself.
For many of our readers, though, the stakes are higher than a personal tool. You may be the product manager deciding what an AI feature should promise, the engineer making it hold up in production for a paying customer, or the founder shipping a first serious AI product who wants to get it right. Work like that needs more than a working demo: it needs evals, guardrails, an eye on the economics, and the judgment to know when a probabilistic system is ready for the people who will depend on it. We built this site to teach exactly that, from the product frameworks behind the decisions to the hands-on curriculum for the build.
This site is a free resource to help you get there.
There is no paywall, no account, no ads, and nothing we are trying to sell you.
Our Privacy and Terms spell out exactly what that means. We collect almost nothing, and we never sell or share it.
Built with AI
Claude Code wrote the code for this site, helped us identify relevant industry examples and research sources, generated the diagrams, explored design systems with us, and drafted nearly all of the prose. We brought the core source material and used Wispr Flow extensively to dictate ideas, feedback, and early drafts whether we were on the move or an idea struck unexpectedly. This helped us capture our thinking while it was fresh and keep Claude working from its latest iteration. Here is exactly who did what.
What Claude Code did
- Wrote the code for every page and feature
- Wrote the course content from our briefs and material
- Drew on its frontier-model training for the drafts
- Researched the industry examples and sources it cites
- Synthesized our research papers with its own sources
- Generated the custom diagrams
- Produced the page layouts and visual design
- Revised everything we sent back
What we did
- Brought our own research, knowledge, and experience
- Dictated it to Claude as the thought files behind the course
- Designed the curriculum and its order
- Set the voice standard and held every sentence to it
- Wrote all the briefs and prompts
- Reviewed, corrected, cut, and redirected the drafts
- Made every call on what shipped
About one in five build prompts was a correction. That tension is the method: use AI for speed, keep people in charge of what “good” means.
We gladly accept the vibe-coder label. Claude worked as our engineer, designer, and writing partner; we supplied product judgment, taste, domain expertise, and years of experience building products people trust. We stayed hands-on throughout: we read every draft, sent back the ones that inflated the stakes or hid who did what, replaced examples that had gone out of date, rejected designs that looked like every other AI-built site, and wrote each correction into the rules file so the next draft started from it.
We also did not invent a private method for the build and then teach you a different one. The Discipline defined the job: turning a probabilistic system into a product that reliably meets a standard. The Human Factors is how we kept our attention and judgment engaged instead of rubber-stamping fluent output. Each page moved through Shape · Ship · Track, and The Operating Manual was the day-to-day playbook, including the fleets of agents we ran in parallel for the largest jobs.
The numbers behind all of this are on How we built this: the prompts we wrote, the corrections we made, the agents we ran, and the hours it took.
Who we are
Hi. We are Shivali and Girish.
Before the titles, we are builders. We are early adopters of AI and deeply curious about how it is reshaping the work we know best, which is building products, setting strategy, and solving real business problems.
Director of Strategy and Growth, Cloudleap
Shivali is a technology and growth leader who works across the full business, from sales and strategy to product implementations and AI adoption. Her work focuses on making AI a meaningful part of products and processes, not something added on the side.
Everything shared here reflects our own views. We build this in our personal time, and nothing on this site represents, is affiliated with, or is endorsed by our employers.
Glad you are here.
This site has two siblings
AI-Native Product Management is one of three things we run. The second is technicalenough.ai (opens in a new window), our free beginner course for people who want only the software foundation, written for complete non-engineers and covering the ground most AI courses assume you already have. The third is FuelTheFam, our family nutrition app and the working product this curriculum gets tested against; its full story is the FuelTheFam case study.
Using this material
The lessons, essays, diagrams, and other material on this site are our work. You are welcome to read it, share it, link to it, and reference it in your own work; if you reuse a substantial passage or a diagram, credit AI-Native Product Management and link back. If you want to use substantial parts of it in a commercial product or paid course, ask us first.
Build with us
Read the Field Notes, subscribe for new ones, or tell us what you are learning by messaging Shivali (opens in a new window) or Girish (opens in a new window) on LinkedIn. We would love to partner with our readers and users to improve this, because a probabilistic system is never finished and we are just getting started.
Free, on Substack. No spam, unsubscribe anytime. Privacy →