Process discovery and governance
Twiggy uses AI to interview the people who do the work, builds structured process models as they explain it, and has every detail validated by the people responsible. One connected process catalogue, from the enterprise overview to each activity.
No sign-up for either. Already have a workspace? Sign in.

Most process documentation is incomplete, out of date, or written by people who don’t do the work. Twiggy starts from the people who do.
Capture how the work is really done, from the people who do it: interview them yourself, or let the AI ask, one question at a time, about what is missing.
Every model goes to the people responsible for it. They confirm what the AI assumed, comment and sign off from a link, with no account.
One catalogue from the enterprise overview to each activity. Hand-offs link one process to the next, so the whole chain hangs together.
Steps link to the systems, risks, controls, roles and requirements that matter, with owners, review dates and a health score that keeps it current.
Why Twiggy, and how it compares
In their own words, to the AI, in a few minutes, on any device.
Who does this? What if it fails? It follows up on what was left out.
Lanes, decisions and hand-offs appear as they answer, laid out automatically.
Anything the AI inferred is marked ? until someone who knows confirms it.
The people responsible confirm or correct it, and approve.
Connected to the map, the next process, and the risks and systems it touches.
See it on the demo bank How it works, step by step The process modelling guide
Traditional process tools help you document processes. Twiggy helps you discover, document and validate them, and keep them current.
| Topic | The traditional approach | With Twiggy |
|---|---|---|
| Access to SMEs | An analyst interviews whichever expert is available, and gets one view of how the process might work. | The analyst stays in control: Twiggy asks the right experts to confirm their part directly, or the AI interviews them, one at a time, whenever suits them. |
| Feedback loop | The analyst writes it up afterwards, some context is lost, and the experts’ time is needed again to review it. | The model is drafted as they talk, and gaps are asked about there and then. Exceptions are captured alongside the flow, and the AI keeps it plain or technical, as the audience needs. |
| Modelling effort | Most of the time and effort goes into making the model look good. | The layout is automatic, so every model reads the same way, and the framework keeps models small by design. |
| Keeping it relevant | Models drift out of date, and people stop trusting them. | Confirming is easy: the people responsible confirm what was assumed and sign off the model itself, and sign-off can renew on a schedule. |
| Processes are hard to find | Processes live in separate documents and drawings. | One connected catalogue: map, chains and flows, linked by their hand-offs. |
| Governance is secondary | Reviews happen when someone remembers. | Review dates, reminders and a health score keep it current. |
Pick a process that matters, interview someone who does it, and have its owner sign it off. Free to start, no card needed.