A platform that already worked, and was ready for what came next.
Doddle originally built the Finbridge Global platform on OutSystems. The MVP went from concept to live in 10 weeks. In the years since, more than 300 organisations have used it and the platform helped Finbridge raise £500k in investment.
The context then changed. The product roadmap had moved toward AI-native workflows, intelligent document processing, retrieval assistants, natural-language analytics. The cost and shape of building those inside the existing low-code estate no longer made sense. Finbridge wanted to move, but without disrupting users, breaking compliance, or running an open-ended programme.
They came back to us for a fixed-scope, fixed-window migration onto a modern AI-native stack, and an honest answer on whether moving was actually the right call.
- 10 weeks
- Original MVP
- 300+
- Organisations on platform
- £500k
- Investment generated
From OutSystems estate to AI-native platform.
AI document processing. Documents land in the platform and come out classified, structured and ready to use, in minutes rather than days. Every extracted field links back to where it came from, so the audit trail is built in.
AI assistant with citations. An in-app assistant that answers questions using the firm's own documents, with citations users can verify and a clear sense of when the platform is confident in its answer.
Natural-language analytics. Ask a question in plain English, get a chart. Permissions are respected, the AI picks from a curated query set, and users can sanity-check the answer.
URL scraping & enrichment. Point the platform at a web page. It pulls the content, tags it against the firm's own categories, and slots it into the right place. Controlled, predictable, no rogue scraping.
AI-powered data enrichment. Records get filled in automatically, extra context, related signals, cross-references. Quality is measured continuously, so accuracy doesn't drift as the underlying AI changes.
Quality testing for AI. Every AI feature is tested continuously. Bad answers are spotted before users see them. Upgrading to a new model is a measured decision, not a leap of faith.
“We started with Doddle on OutSystems, fast MVP, fast iteration. As we matured, they helped us move onto a full AI-native stack. The speed, the thinking, and the care in execution have been exceptional.”
A platform the team can grow into.
The application went live on the new stack on schedule. Users experienced the move as a UX refresh rather than a re-platform. Run cost dropped meaningfully. The AI features that had been the original motivation shipped iteratively in the months after launch, each one shippable independently, each measured.
The team can now do things on the platform that weren't possible before. They'll also be the first to tell you that the result wasn't a verdict on OutSystems, it was the right answer for them, in their context, at this moment.
