Map a legacy system without source code

A Thoughtworks team used Gemini to build and check functional blueprints for old systems when source code was unavailable.

Done with Gemini· Intermediate

What they did
Thoughtworks consultants ran a two-week pilot with Gemini 2.5 Pro to help unpack legacy systems that had no source code on hand. The model was used to identify what the systems did, then the team validated the results as blueprints of system behavior.
How it went
The pilot produced validated blueprints of the systems’ functionality in two weeks. The takeaway was that AI could speed up modernization work for hard-to-read systems, but only if the output is carefully checked.
Tips
Treat the first pass as a draft, not a final answer. Build in a review step so you can confirm anything the agent labels uncertain before you rely on it.

Try it yourself with Gemini

You are helping me understand a legacy application at [system/location] where source code is not available. Review any documentation, config files, logs, database schemas, and related artifacts I provide, then draft a markdown blueprint that explains the system’s purpose, main workflows, data movement, and business rules. Separate confirmed findings from anything uncertain, and flag any assumptions clearly. Stop after you finish the blueprint and wait for my approval before making any changes, sending anything, deleting anything, or submitting anything.

Fill in the [brackets]. Agents can make mistakes, so review anything before it is sent, paid or deleted.

Read the original on infoq.com ↗

Get 3 agent uses like this every morning

With the day's top AI news. Free, and one click to unsubscribe.