Docs

Help desk / Measure

Insights (Experimental)

See which escalated requests Risotto could automate next, act on them with the Assistant, and find where your knowledge has gaps

The Insights page shows what your tickets say you could automate next, and where your knowledge has gaps. Open it from Insights in the sidebar, below Analytics. The page has two tabs: Automation opportunities, which opens first, and Knowledge gaps.

Who can see Insights

Insights is for Org Admins and Org Managers who are a member (Agent or Supervisor) of at least one department. The page shows only the tickets of your own departments and their sub-departments. Being an Org Admin does not add other departments.

Anyone else does not see Insights in the sidebar, and opening the page's URL takes them to Analytics.

Automation opportunities

The Automation opportunities tab covers the last 90 days. Risotto calculates it when you open the page.

The card at the top shows two figures and a bar:

  • The share of tickets that Risotto autosolved, for example "44.2% autosolved (548 of 1,240 tickets in the last 90 days)". This rate divides by all tickets, so it is lower than Solved automatically on Analytics, which divides by solved tickets.
  • How many of the classified tickets are automatable. Classified tickets are the escalated and human-resolved ones. When a ticket escalates, Risotto classifies it and names the capability that could have resolved it: a tool you already have, a tool template, an API source, an access rule, a runbook template, or a new integration.
  • A bar that splits all the tickets into Autosolved, Automatable today, Automatable with setup, Needs a custom tool, No automation path found, and Not classified. Each ticket is in one band, by the hardest capability it needs. Point to a band name with a dotted underline to see which filters of the table match it.

The table Tickets that cite an opportunity, last 90 days has one row for each capability, with the number of tickets that cite it. A ticket can cite more than one capability, so do not add the rows up. The table shows the top 40 rows, and a line under the table says how many more were left out. Access rule and runbook template rows always rank first.

Each row shows:

  • Opportunity: the capability, for example Access: Salesforce or Tool: Create Slack Channel. The line under it adds details when they apply: disabled, needs a runbook, the template it comes from, the resources the tickets requested, or how many spellings of the name the tickets used.
  • Type: the kind of work it needs: Create access rule, Use runbook template, Did not auto-resolve (the tool is ready, but the tickets still reached a person), Finish setup, Configure template, Promote source, Build integration, or Deleted (the tool was deleted since).
  • Tickets: the tickets that cite it.
  • Action: the next step.

The filters above the table have one button per type, with the number of distinct tickets of that type. Click a filter to show only that type, and click it again to show all rows. A filter can show Nothing under this filter when all of its rows fell outside the top 40. Its tickets still count in the total. Click Show all opportunities to clear it.

Act on an opportunity

  1. Filter the table

    Optionally, click the type you want to work on, for example Create access rule. The table shows only those rows.

  2. Click the row's action

    Click the action in the Action column.

    • Enable tool, Add to a runbook, and Open tool open the tool.
    • Create access rule, Use template, Ask why, Configure template, Promote source, and Ask the Assistant open the Assistant panel on a new conversation, with a prompt for that row in the message box.
    • A - means there is nothing to open, for example because the tool was deleted.
  3. Send the prompt

    The Assistant does not send the prompt until you do. Edit it if you want, then send it. Your earlier conversation stays in the Assistant's history.

Knowledge gaps

The Knowledge gaps tab lists two kinds of finding:

  • Missing knowledge: a theme of at least 3 related tickets from the last 90 days that Risotto could not answer, for example tickets that escalated or got a thumbs-down. The row suggests a document to update or a new one to write.
  • Stale suspect: a document that Risotto cited on tickets that ended badly. It may be out of date. Check it before you change it.

The heading says when the findings are from, for example Gaps the analysis found, 2 days ago. The filters under it show how many findings of each kind there are; click one to show only that kind. Each row shows:

  • Gap: for a stale suspect, the connector of the document (for example confluence), then a summary, why it is a gap, and what to do about it.
  • Type: Missing knowledge or Stale suspect.
  • Tickets: the tickets behind the finding. A ticket can be behind more than one finding, so do not add the rows up.
  • Action: Draft the article opens the Assistant panel on a new conversation with a prompt to draft content for the gap. As on the other tab, the prompt is not sent until you send it. Open document opens the cited document in a new tab. A - means the document has no link.

An escalation counts toward a finding only when Risotto escalated because its knowledge did not answer the request. A hand-off for another reason, for example an action that only an admin can do, a policy decision, an outage, or a tool failure, does not create a Missing knowledge theme and does not mark a cited document as Stale suspect. Escalations by your team, and older escalations, still count.

Run the analysis

Risotto runs the analysis when you open the page and there is no run yet, or the last run finished more than 7 days ago. Your organization runs one analysis at a time. To run it now:

  1. Open the Knowledge gaps tab

    Click Knowledge gaps at the top of the page.

  2. Start the analysis

    Click Run analysis. Risotto shows "Analysis started. The findings update on their own." If the run cannot start, Risotto shows "Could not start the analysis. Try again in a moment."

  3. Wait for the findings

    While the run is in progress, the button shows Analyzing… and the page keeps the findings of the last complete run. The findings update on their own when the run ends, so you don't need to reload the page.

If the run fails, the page says The last analysis failed and keeps the findings of the run before it. Click Run analysis to try again. Before the first run, the tab shows Not analyzed yet. A run that finds nothing shows No knowledge gaps found.

Something wrong or missing? Report an issue · View as Markdown