The start: analysis instead of a specification
We don't start with a forty-page document. First we look at what you already have: spreadsheets, the current system, e-mails, exports. For FoxFix that was a technical audit of the original system. We went through its code and database and only then recommended rebuilding rather than patching it again. For Predictimo we first used a trial model to check that the data produced meaningful results at all.
- We map where routine work piles up and where data gets retyped.
- We pick one process, for example the path from enquiry to quote to invoice, and the prototype covers it end to end.
- We agree on how we will know the prototype has succeeded.
Next: a prototype on your data
A prototype isn't a clickable picture. It is a working app that covers one full process from start to finish. Wherever possible it uses your real data from spreadsheets or the existing system. The ZEUS Core mobile app, for instance, first ran on sample data, and we connected it to the live system in the next phase. We polish the interface only where a decision depends on it. The rest is deliberately rough.
You see progress as it happens, not at the end. We make most decisions together at regular meetings and check-ins over the running app. No specification can replace that.
What you get and what gets decided
At the end of the prototype you have three things: a running prototype, a realistic estimate of the scope and price of the next phases and, above all, an answer to whether it is worth continuing. Sometimes the answer is “yes, but differently than we thought”. That is exactly when the prototype has saved the most money. If we continue, the prototype isn't thrown away. It becomes the foundation of the live version.
Why it works for the budget too
A prototype is the cheapest way to learn the real scope. An estimate after the first piece of real work is far more accurate than one from a document. By then we know what your data looks like, which integrations are truly needed and what people actually use. The prototype costs a fraction of the total budget, and the risk of building the wrong thing drops sharply.