Investigate the gaps.
Map the data path, model calls, permissions and dependencies. Agree the actual requirements and success criteria.
Prototype to production
The demo works.Now make it dependable.
Review an existing prototype and close the gaps that matter: identity, integration, testing, monitoring and operational ownership.
Explore the capabilityENGINEERING FOCUS
01Assess the prototype
02Close the critical gaps
03Validate and deploy
Not everything needs rebuilding.
Understand what exists before prescribing a replacement. The useful work may be a stronger interface, better tests or one missing control, rather than a new technology stack.
Engineer the foundations.
Fix the critical failure modes, automate deployment and add appropriate controls. Make the behaviour observable.
Test the operating model.
Validate the workflow with representative users and test recovery. Agree ownership, documentation and the route into support.
A possible workflowIllustrative example.
Retain a working retrieval prototype while rebuilding its permission checks and adding a regression test suite.
A prototype demonstrates an idea, not all the requirements of a production service. Scope the missing work before promising a release date.
A closer look
Good questions.
Straight answers.
Can you review code another team wrote?
Yes. Access, documentation and a clear review scope help establish the starting point.
Do we have to start again?
Not by default. Preserve useful work and explain any proposed replacement against the risks and requirements.
Technical reference: Anthropic: evaluating AI agents (opens in a new tab)
Keep exploring