Alpha Prototype

The problem to solve

Ideas get discussed. Workshops get run. But how does any of it become something real? RAND found more than 80% of AI projects never reach meaningful deployment, most often because leadership and the team never agreed on the specific problem being solved.¹ These two weeks are built so that doesn't happen here.

The format

The alpha prototype is grounded in the formats we used to ship AI features at Xero.

It starts with Discovery: understanding the workflow, the jobs to be done, who it serves, and a baseline of the underlying tech. From there we lock scope and define "success," before moving into concept development and building the alpha itself. Once it's working, we put it in the hands of real users, then run one round of refinement addressing what testing showed. We close with a go or no-go decision: keep it, iterate, or stop with your eyes open.

The outcome

You're left with a functioning, interactive artefact and feedback from real users, enough to make a real go or no-go decision, not a vague "let's keep exploring." Just as crucial: a repeatable process framework, ready to evolve the current prototype or apply to the next problem.

Past alphas: a project management tool for residential builds, a cashflow forecasting tool, and a rostering solution for a tech consultancy.

Out of scope: backend build and backend integration for the alpha sprint.

Prerequisite

Access to the workflow owners, users for testing, and feedback throughout. That's it.

This can run as a facilitated sprint your team picks up afterward, or independently once your team has the framework.

Investment

From $3,000 plus GST for a two-week sprint.