Practice 04 · Labs & Prototyping

Test the idea.
Choose the next step.

A focused sprint to turn uncertainty into evidence.

Discuss a sprint

One question. A working experiment.

Illustrative sprint

The question

Can enquiries reach the right team?

The assumption

Routing can be made useful with topic and completeness checks.

What we looked at

  • Matching ruleTopic keywords map an enquiry to one of three teams: Sales, Operations or Support.
  • Missing-information flagFour named fields are checked before any routing is attempted.
  • Human review for edge casesWhen the rules cannot pick one team, a person decides. The prototype never guesses.

Fictional enquiries and generic teams. The prototype runs in your browser; nothing is sent.

Example sprint sequence

Two weeks, four phases.

  1. 01DefineFrame one question and agree what would count as evidence.
  2. 02MakeBuild the smallest prototype that can test it.
  3. 03TestRun it on real or representative inputs and record what happens.
  4. 04ReviewRead the evidence together and choose the next step.

Some questions close with a clear answer. Others close with a narrower question, and that is recorded as well.

Build, refine or stop. Each one is an answer.

Every sprint closes with one of these three decisions and the evidence behind it.

  • Build

    Evidence supports a next build.

    The build is scoped and priced on what the sprint showed.

  • Refine

    Resolve a remaining uncertainty.

    A narrower test, or a change of approach, before anything larger.

  • Stop

    Do not pursue this approach on current evidence.

    You keep the code and the write-up. The budget stays free for something else.

What you keep.

A clear record of what was tested, learned and decided.

You keep the code and the write-up whatever the recommendation. A recommendation to stop is a useful result too: it costs two weeks and CHF 3,500 instead of a full build.

Internal R&D

Internal R&D · not client work

Morvion AI Sentinel

An operating-system concept for running a business day to day: CRM intelligence, tasks, multi-agent workflows, sales pipelines and decision support. We ran it as a self-funded Discovery Sprint. A Morvion product direction, not a client engagement.

See studio R&D

Questions

A prototype tests one risk: feasibility, user response or an integration. An MVP is the smallest version real users can rely on. When a risk is open, we run the prototype first, in two weeks, and build the MVP only when the evidence supports it. A focused MVP can be ready in two weeks. A scoped product can launch in around two months. We agree the timeline after reviewing your requirements.

Deployed code, a one-page architecture record, the test findings (with the evaluation data for AI sprints), a written recommendation, a cost and timeline for the next phase when the answer is build, and a review call. No slide deck.

Not always. A sprint tests one assumption. If the evidence is not conclusive, the recommendation says so and names what is still uncertain.

A fixed price of CHF 3,500, paid upfront. The Discovery Sprint (the Digital Diagnostic on our price list) tests one assumption, so its scope stays small enough for one fixed price.

Bring the assumption that worries you most.

One question is enough to scope a sprint. We will tell you on the first call whether a sprint can test it.

Discuss a sprint