Indicative scope: 3–6 weeks
Test the assumption before scaling the work.
A prototype is useful when a specific question remains open, such as whether a workflow can be supported, whether sources are sufficient, or whether people can review the output effectively.
Who it is for
Teams with a defined use case and access to representative examples, who need technical or user evidence before deciding on a fuller implementation.
What a prototype can clarify
- Can the proposed approach handle the cases people actually encounter?
- Where does it fail, and what review or fallback is needed?
- Do the expected benefits justify the integration, operating and maintenance effort?
Build around a testable question
The work begins by setting a narrow scope and agreeing representative scenarios. Good1 builds the smallest working prototype that can answer the question, then evaluates output quality and technical feasibility against those examples. The findings include failure cases, dependencies and operating effort.
What may be built
- A document-processing flow with a person reviewing extracted fields.
- A retrieval-based assistant that answers from an agreed document set and shows its sources.
- A contained AI feature connected to a test environment or representative product flow.
What you receive
- A working prototype within the agreed test boundary.
- Evaluation notes tied to representative scenarios, including known failure modes.
- A recommendation to proceed, revise the approach or stop, with dependencies for the next stage.