The risks were said out loud,your RAID log heard them
Every steering meeting surfaces a risk, an assumption nobody wrote down, a dependency mentioned in passing. Khint turns the raw notes into clean RAID log entries with one AI action, keeps the project's context in a Memory session, and publishes the log to Confluence, so the register stays alive between meetings instead of being rebuilt before the audit.
RAID logs don't fail in the template,
they fail in the follow-through
Everyone agrees the project needs a risk register, and every project starts one. Then reality: risks get raised verbally in a steering meeting, land as a half-sentence in someone's notes, and the log is updated “later”, which means the night before the review. The gap is not knowing what a RAID log is; it is the ten minutes of reformatting between messy notes and a presentable entry, times every meeting. Khint removes exactly that gap: one saved Actionconverts the notes you already took into entries in your team's format, and the palette pushes the result to the Confluence page where the log actually lives.
From meeting murmur to published register
One action, one session, one shortcut: the log keeps itself current.
Save a “Notes → RAID entries” action, once
In Agents, create an Action whose prompt describes your team's RAID shape: category, entry, owner, next step. No prompt yet? Type a one-line description and press Generate; tighten the columns later with Improve. From now on the log has one format, not one per meeting.
Start a Memory session for the project
Open Memoryand start a session named after the initiative. Pull the epic or the active sprint from Jira through the palette, and that context is compacted and carried into every action you run: the model reads “vendor API” in your notes and already knows which vendor and which deadline you mean.
Select the raw notes, hit the shortcut
After the steering meeting, select your messy notes wherever they are (the meeting doc, an email, a text file), press Cmd+Shift+K, and run the action. The worry someone voiced becomes a risk with an owner; the “we're waiting on the data team” aside becomes a dependency. Clean entries are pasted back exactly where your cursor was.
Publish the log to Confluence
RAID logs die in private notes. With Atlassian connected on the MCP page, the palette's Confluence → Create page row turns the updated entries into a page in your space, and the write is logged into the session. The log lives where stakeholders already look, not in a file only you can find.
Escalate a risk the moment it turns real
When a risk becomes an issue that needs work, run Jira → Create issue on that entry and it becomes a tracked ticket. Do this often? Chain the extraction action and the Confluence step into a Workflow, and the whole notes-to-published-log routine becomes a single keystroke.
What an extracted entry looks like
Four kinds of entry, pulled from the same page of notes.
The classification comes from your notes, not from thin air: the action reads what was actually said and sorts it into your categories. Because the active Memory sessioncarries the compacted history of the project (the pulled epic, the previous entries, the last status), a vague “the vendor thing is still shaky” resolves to the right vendor and the right deadline. Compare that with a status report: the report is a snapshot for stakeholders, while the RAID log is the running ledger the report draws from. Khint maintains both from the same notes.
Common questions
What is a RAID log, and how does Khint structure one?
A RAID log tracks a project's Risks, Assumptions, Issues, and Dependencies in one place, so nothing surfaced in a meeting silently disappears. In Khint you save an Action whose prompt describes exactly the entry shape your team uses (category, description, owner, next step), then run it on your raw meeting notes. Because the prompt is yours, the output matches your template instead of a generic one.
Does the AI invent risks I never discussed?
No. An Action only sees the text you selected, plus (if you opted in) the compacted context of your active Memory session. It classifies and rewrites what is already in your notes: a worry someone voiced becomes a risk entry, a blocked hand-off becomes a dependency. You always see the result before it lands anywhere: it is pasted back where you are working, or shown to you before a Confluence page or Jira issue is created.
Can Khint put the RAID log where the team expects it?
Yes. Connect Atlassian from the MCP page with your email, API token, and cloud domain, and the palette gains Confluence Create page and Jira Create issue rows. Publish the updated log as a Confluence page, and when a risk turns real, raise it as a Jira issue without leaving your notes. Each write is also logged into your active Memory session, so the project's history stays in one thread.
Do I need to write the extraction prompt myself?
Only if you want to. In the Action editor, type a one-line description like "extract risks, assumptions, issues and dependencies from these notes as a RAID table" and press Generate: Khint drafts the prompt for you. Press Improve later to refine it in place, for example to add an owner column or your severity wording. The free plan includes 10 AI actions per day, no credit card.
Keep the register current, not reconstructed
Free with 10 AI actions and 5 captures per day. No credit card. Save the action once; the log updates itself from your notes.