End-to-end ownership. Initial discovery, planning and scoping, designing and iterating in Figma and Paper, developing and shipping the front-end.
Anatomy of a RAID log
A RAID log is a collection of Risks, Action Items, Issues, and Decisions. It documents the status of a project as it evolves, identifying threats to the critical path and ensuring appropriate ownership and traceability.
In digital transformation programs, where projects span three to five years, cross project dependencies, large teams, and even larger budgets make a RAID log critical to success and extremely difficult to manage.
RAIDs crop up in meetings, chat messages, emails, reports, and any other surface where people interact. They get discussed and tackled in an even broader set of tools, and their evolution is long and rarely clear-cut. This complexity is an opportunity for Tato to put its deep project knowledge, integrations, and agents to work to extract, monitor, and document RAID items across the lifecycle of a multi-year project.
Extraction and human-in-the-loop
Meetings are the primary source of RAID items, but the items themselves are usually managed in Jira, Smartsheet, or plain old spreadsheets. Our first job was consolidating those sources into a normalized log that could act as a single source of truth. Tato updates the log after any activity, and pulls from external tools periodically or on command through MCPs and bespoke integrations.
The RAID updates on the activity summary page started as a simple section showing the under-the-hood thinking of our agent. Exposing that machinery is almost always a mistake that overloads the user, and despite our best efforts, LLMs produce volume, which leads to an overwhelming number of items.
Hiding the machinery was the first improvement, and it cut a lot of noise on its own, since multiple snippets often related to a single update on a single item. From there I reduced the amount of text and added icons and avatars, which improved legibility and hierarchy and made the whole list easier to parse.
Risks
Action Items
Issues
Decisions
I also exposed two primary actions on each item, so tracking or dismissing is always within reach and the log stays tidy.
Dismissing an item tucks it away, letting the system know this AI extraction was not useful. Tracking an item marks as exceptionally relevant, so I used a subtle interaction to make it feel impactful and energetic.
This first iteration landed well, with early feedback showing a perception that Tato was smarter and more nuanced. We've found our users' quantitative perception to be just as important as the qualitative improvements when it comes to interacting with an LLM.
To take it further I reworked the summary page from a single vertical column into two columns, giving RAID management its own first class surface where the user keeps meeting context in view while digging into items for proper grooming.
Risks
2Action Items
3Issues
1Decisions
1An auditable paper trail
A well managed RAID log is also an accurate record of a project's evolution, and of what the team did to mitigate critical risks. Our history view was hard to reach, sometimes hidden behind two separate clicks, and it focused on diffs between versions. Diffs matter, but they fail to paint the most important picture: what caused a change, and when.
Clarity matters even more when working alongside an agent. The history documents which changes the agent made and the reasoning behind them, so the human can understand them and correct them when necessary.
Risks
3Action Items
4Issues
2Decisions
2Recorder can't join authenticated meetings
The recorder is turned away from any session that requires a tenant sign-in, so a workshop run on the customer's own account produces no transcript and nothing reaches the log until someone writes it up by hand. Two of the last four workshops were run that way, and the items from both were reconstructed after the fact.
I tried many iterations for showing an item's history and ultimately kept the drawer we already had. The drawer offers the most real estate for understanding a risk and its history while letting the user stay in their current workflow, moving easily between every item in a list. Making it more visual and tightening the hierarchy also gave space back and increased legibility.
The ultimate control centre
The last piece of this iteration was an improved control centre. Much of the work happens in the context of an activity summary, and the components and flows are largely shared, but we still wanted a delightful surface for managing the full log.
I started with basic UX improvements: tightening space for more breathing room, grouping related controls, and making the table contents scrollable rather than the page, so the header, footer, and every available action stay visible.
Along with all of the other status icons to show impact and likelihood, I built this Linear inspired fractional progress indicator so we can smoothly animate between the states, and support any number of intermediate states.
Finally, based on user feedback, I made the table editable without opening an item. That feedback partly inspired the drawer redesign too, making every property of an item visual as well as editable inline.
The future of RAID
The goal of this project, and of Tato as a whole, is to make the processes behind these large transformations more automated and effortless, shifting the human workload from note taking and data transfer to verification and the occasional correction.
This RAID log is one small step toward a future where humans and AI collaborate to ensure successful business transformations.
Next case study
Customer knowledge base