Case study · Custom AI Systems
Structuring a product before a line was built
A healthcare and finance company came looking for someone to start building. Stallwart started somewhere else: with the questions that had to be answered before anything was built.
- 01The hard questions first
- 02Structure the product
- 03A plan the team could build
- Client
- a healthcare and finance company
- Problem
- A product ready to be built, before the questions underneath it had been settled.
- Approach
- Ask the hard questions first and structure the product before any code.
The problem
The client came to Stallwart expecting the usual: a team ready to start building the company's product. He had sat through plenty of calls with people hired to sound smart, and had learned to expect a pitch and a deck.
The problem was not a shortage of people willing to build. It was that the questions underneath the product, the ones easy to avoid, had not been asked out loud yet.
The reality behind the problem
Hiring a team and starting to build looks like progress because it produces motion. But motion on an unsettled product just builds the wrong thing faster. Bringing in people to build defers the questions underneath the product rather than answering them, and the questions do not wait.
The questions easy to avoid are the ones that decide whether the product holds together: who it is for, what has to be true before it is built, where its boundaries sit. Left unasked, they do not disappear; they surface mid-build as rework. The real work was not more hands, it was facing those questions first.
What was assumed
We need someone to start building.
What Stallwart asked
What has to be true before any of this is built?
The desired outcome
The goal was not a faster start. It was a product whose hard questions were answered before anyone wrote code, so the build had something solid to stand on.
- The questions underneath the product named out loud, not avoided.
- Each decision owned by the founder, settled by him rather than sold to him.
- The scope and shape of the product clear enough to build against.
- The order of the build settled, so the team knew what came first and why.
- A plan the team could act on without discovering gaps mid-build.
The system we designed
Stallwart did not start by building. On the calls, Stallwart asked questions, the real ones the founder had been quietly avoiding, and then stayed with them while he worked through the answers. No pitch, no deck. What came out was the shape of the product: the decisions that had to be settled before any code, structured into a plan the team could build from.
How it works
- 01
Ask the hard questions first
Not the comfortable ones. The questions the founder had been avoiding came first, because the product could not be structured around answers nobody had faced.
- 02
Sit with the answers
No pitch, no next deck. Stallwart stayed on the call while the answers got worked through, so the decisions were his, settled, not sold to him.
- 03
Structure the product
The settled decisions became the shape of the product: scope and structure clear enough that the team knew what to build and in what order.
The approach
Structure before a line was built.
- 01
The hard questions first
The ones the team had been avoiding.
- 02
Structure the product
Decisions settled before any code.
- 03
A plan the team could build
Scope and shape, clear.
No pitch, no deck. The questions that had been quietly avoided, worked through until the shape of the product was clear.
Architecture
What sits underneath.
What the engagement structured was not a codebase but the product itself: the decisions and boundaries settled before any build, laid out as layers the team could build from.
The questions underneath
The hard questions the founder had been avoiding, named and worked through first, because nothing above them could be settled until they were.
The decisions settled
Each open question turned into a decision the founder owned, so the product rested on choices that had been faced rather than on assumptions nobody had tested.
The shape of the product
The settled decisions arranged into scope and structure: what the product was, where its boundaries sat, and what it did not try to be.
The build plan
The structure turned into an order of work the team could follow, so the first line of code had a settled direction behind it.
The hard parts
Where the engineering judgment was.
Questions before building
A product built before its hard questions are answered is built on guesses.
The trade-off
Slower to the first line of code, in exchange for a product that does not need rebuilding.
No pitch, no deck
The goal was clarity for the client, not a sale for Stallwart.
The trade-off
Nothing to hand over at the end of the call except a decision the founder could act on.
Structure, not features
The engagement was about the shape of the product, settled before code, not features rushed out.
The trade-off
No visible build yet, but a plan solid enough to build against.
Reliability & guardrails
How it avoids the wrong call.
Reliability here was about the product not failing in the ways that are expensive to fix later: being built on guesses, or rebuilt once a question finally gets asked.
Questions faced, not deferred
A question avoided at the start becomes rework mid-build. Facing them first removed the failure that begins with an assumption nobody tested.
Decisions the founder owned
Because each decision was his, settled rather than sold, the product did not rest on a direction he would later reverse.
Boundaries made explicit
Naming what the product was and was not kept scope from drifting once the build started, where drift is what turns a plan into a rebuild.
A plan that could be checked
The structure was concrete enough that a gap would show up on the plan, before it showed up as code that had to be thrown away.
Productionization
What turns it from a demo into software.
What kept the engagement from being a good conversation that faded: turning the clarity into something the team could build from and return to.
Decisions written down
The settled decisions were captured, so the clarity survived past the call and the team could point to a reason behind each choice.
Scope and sequence documented
The shape of the product and the order of the build were laid out, so the plan could be handed to the people who would build it.
Clear enough to act on
The founder left the calls knowing what to do, which is the test of whether structure is real or only discussed.
A direction the team could check against
With the product's shape settled, later decisions had something to measure against instead of reopening the same questions.
What changed
What changed was not a feature set but the ground the product stood on. The founder came in expecting a pitch and a team to start building; he left with the hard questions answered and the decisions settled, knowing what to do next.
The product had a shape before it had code. The team could build from a plan rather than from guesses, and the choices underneath it were the founder's own, faced rather than deferred.
- From hiring people to build toward answering what had to be true before building.
- From questions quietly avoided toward decisions faced and settled.
- From a pitch and a deck toward clarity the founder could act on.
- From a product built on guesses toward one structured before any code.
In practice
Founder, a healthcare and finance company
The founder said he had paid many people to sound smart on calls with him, while Stallwart simply asked questions: the real ones he had been quietly avoiding, and then stayed with him while he worked through them. There was no pitch and no next deck, and he got off the call knowing what to do.
Ownership & handover
What you receive, and keep.
What the founder kept from the engagement, his to build on, with nothing held back.
- The hard questions, named and answered.
- The decisions underneath the product, settled and owned by the founder.
- The shape of the product: its scope, structure, and boundaries.
- The order of the build, so the team knew what came first.
- The clarity to act, independent of Stallwart.
What we learned
The insight that generalizes: when someone asks for a team to start building, the useful first move is often not to build but to ask what has to be true before anything is built. A product structured before code rests on decisions that have been faced, not on guesses that surface later as rework. The questions easy to avoid are usually the ones that decide whether the product holds together, and facing them first is the cheapest work in the whole engagement.
This work connects to
- Custom AI Systems
- Product architecture
- Regulated industries
Frequently asked
What is healthcare product architecture?
It is the work of structuring a product, the decisions, scope, and shape, before any code is written. For the company it meant answering the hard questions underneath the product first, so the build had something solid to stand on.
Why structure a product before building it?
Because a product built before its hard questions are answered is built on guesses. Settling the decisions first means the team knows what to build and why, rather than discovering gaps mid-build.
Did Stallwart build the product on this engagement?
This engagement was about structuring the product before building. The work was the questions and the decisions underneath them, turned into a plan the team could build from, not features shipped.
What did the company get out of it?
Clarity. The founder left the calls knowing what to do: the hard questions answered, the decisions settled, and the shape of the product clear before any code.
Not sure what has to be true before you build?
We ask the questions you've been avoiding, and leave you knowing what to do.
Last updated: October 7, 2026