My First Big Ask

On June 6, 2026, I sent the first big prompt for MyUPP.

And when I say big, I do not mean, “Please make me a nice little form.”

I asked for the beginning of a Universal Provider Passport — a national physician credentialing platform. The idea is simple to say and very hard to build: help reduce physician credentialing from the usual 90 to 180 days down to days.

Credentialing, for anyone lucky enough not to know the word, is the process of proving a medical provider is who they say they are, has the licenses they say they have, and is approved to work with hospitals or payers. It matters enormously. It is also slow, repetitive, and full of paperwork.

So naturally, as a 50-year-old non-technical person, I decided to ask an AI app builder to help me take a swing at it.

No small plans over here. Apparently I like to learn by standing at the bottom of a mountain and saying, “Well, that looks interesting.”

In that first prompt, I laid out the big vision: a platform with different portals for providers, hospitals, payers, and credentialing specialists. I wanted the provider to have a passport — one place where their professional information and documents could live. I also wanted a credentialing engine, meaning the part of the system that helps track what is complete, what is missing, and what still needs review.

I knew the full version was too much for a first build, so I asked for a V1. That became the Provider Passport plus the Credentialing Engine. I also asked for a serious, healthcare-grade visual feel. The words I used for the look were along the lines of clinical trust: deep navy, medical teal, clean, calm, not flashy.

I also asked for real foundations, not just a pretend demo. Login by email and password. Google sign-in. A real backend using Supabase, which is basically the behind-the-scenes place where the app stores information. I wanted role-based access too — the idea that different people should only see what they are allowed to see. In the technical world, one piece of that is called RLS, but in plain English it means guardrails around the data.

What came back surprised me.

It did not come back as a complete national credentialing system, of course. That would have been magic, and also probably dangerous. But it did come back with the shape of something real.

There was a multi-portal shell. There were places for different user types to exist. There was the beginning of authentication. There was a real database connection. There was a design direction that actually felt like healthcare instead of a random tech template wearing a stethoscope.

For a few minutes, I felt wildly encouraged.

Then I looked at the scope again and felt my stomach drop.

Provider portal. Hospital portal. Payer portal. Credentialing specialist portal. Documents. Reviews. Statuses. Security. Trust. And eventually, maybe document scanning and a marketing landing page.

I remember thinking, “This is either the smartest way I have ever learned something, or the most elaborate way to embarrass myself.” Maybe both.

Did it live up to expectations?

Yes, but not in the way I expected.

I think I secretly hoped the AI would turn my big idea into a nearly finished product. What it actually did was more useful: it gave me a first version I could react to. It made the invisible visible. It turned a giant foggy idea into screens, structure, and decisions.

That is when I realized the first prompt is not really about getting the final answer. It is about getting something on the table.

Something to question.
Something to simplify.
Something to argue with.
Something to improve.

And maybe that is the part I needed most. Not perfection. Just a beginning that was good enough to make me keep going.

Small takeaway: A big prompt will not finish the mountain, but it can show you where the trail starts.