Case Study Identification with AI Case Study

The Business Problem

Every client system has years of history behind it: hundreds of issues, thousands of code changes and the conversations around them. The work that would interest a new client, such as a calculation fixed, a manual task automated or a new product launched, is buried among routine changes. Reading all of it by hand to find the best stories would take weeks, so our website showed only a handful of older case studies.

The Solution

We set up a repeatable process for each client. AI reads through the client's issues and code history for a chosen period, groups related changes into features, and suggests which ones have a clear problem, solution and result that a business owner would recognise. It also counts the work for the client's Development Highlights, such as improvements delivered, new features and the size of the system.

Clear rules keep it safe. Nothing in the history is published as it stands: issue numbers, code, staff and customer names, and anything sensitive are kept out, and only clients and partners who have agreed can be named. We check every story against the real work before it goes live, and a log records which work has been covered, so no story is told twice and the next run starts where the last one ended.

The new case studies section is built straight from these studies, with each client's overview, their highlights and a page for each of their systems.

The Result

In a few days the site went from a handful of case studies to more than 90, covering seven clients and their systems. Each client's history can now be revisited whenever new work is finished, and visitors can see the depth of what we do.

See all case studies for Website