# Every escalation tells you what to fix. Risotto reads them all.

Risotto Insights runs AI agents in the background. They read every request your team had to handle and tell you what to automate, which docs to write and which to fix. Hand a finding to the Assistant, approve the change, and Risotto handles that request next time.

[Book a demo](/demo)

> "One special thing about Risotto is it's constantly improving with every question." — Jason Huey, Senior IT Systems Administrator, ThoughtSpot

A finding in Insights becomes an access rule the Assistant drafts for your approval. Example data.

## How Risotto gets better every week

1. **Find.** Every time a request goes to a person, an AI agent reads it and works out what would have resolved it.
2. **Rank.** Insights adds up the last 90 days and puts the fixes that cover the most tickets first.
3. **Fix.** Hand a finding to the Assistant. It builds the access rule, runbook or article, and you approve the change.
4. **Learn.** When your team answers a question in Slack, a ✍️ reaction turns the thread into a knowledge article.

## Know exactly what to automate next

Your escalations already say what to automate. Insights reads them for you, checks each finding against your real tools and runbooks, and turns the result into a short list:

- **Apps people keep asking for.** Create the access rule, and the next request goes to approval and access, not to the queue.
- **Requests a template already covers.** Start from a runbook template that's ready to use.
- **Tools you set up but never turned on.** Open the tool from the row and finish the setup.
- **Systems your tickets keep naming.** Plan the integration with the Assistant.
- **Automations that didn't finish the job.** Ask the Assistant why, and fix it.

A bar at the top shows how many tickets Risotto solved, and how many more it could solve today or with more setup.

Insights for the last 90 days. Example data.

## Your knowledge base goes out of date. Insights notices.

A VPN client changes, a policy moves, and the old article keeps sending people the wrong way. A second AI agent reads 90 days of tickets for answers that fell short: escalations with no answer, searches that found nothing, and answers rated thumbs-down. It turns them into a to-do list.

- **Questions no article answers,** with how many people asked and what to write.
- **Outdated docs** that keep leading to escalations, with a link to open them.
- **A first draft from the Assistant,** started from the finding, for you to review.

The knowledge gap report. Example data.

## Ask any question about your support

Dashboards answer the questions someone thought of in advance. For the rest, ask the Assistant in plain English. It writes a short program with Code Mode, runs it on your support data, and answers with the numbers and a file you can download.

- "Which request categories grew the most this quarter?"
- "Export every open ticket with its assignee to a CSV."
- "List Workday hires from the last 90 days by department."
- "Build me a dashboard of this month's automation opportunities."

It reads tickets, comments, people and both Insights reports, plus the tools you connected. A program that only reads runs at once. Anything that would change something waits for your approval, and every call is logged.

Code Mode answers a question about tickets. Example data.

## Show leadership what Risotto saved

Analytics keeps AI work and human work apart, so the numbers hold up in a budget review.

- **Time saved,** worked out from your own AI and human resolution times.
- **Solved automatically,** counted only when Risotto finished the request with no person stepping in.
- **AI and human time to resolve,** side by side.
- **A Year in Review report** of the past 12 months, ready to share.

Filter by department, assignee and date range, and select any percentage to see the tickets behind it.

The analytics dashboard. Example data.

## What customers measure

- **6.5 hours** average AI resolution time at ThoughtSpot, against 31 hours for human agents. [Read the story](/customers/thoughtspot-automates-multi-department-support-with-ai)
- **3 hours** to resolve at Jobber when Risotto handles a request, against 35 hours with a person. [Read the story](/customers/jobber-builds-enterprise-wide-ai-support-funnel-with-risotto)
- **Under 1 day** to resolve at Retool, against a 2-day SLA before. [Read the story](/customers/retool-sla-customer-story)

## Frequently asked questions

### How does Risotto find what to automate next?

Risotto Insights uses an AI agent that reads each escalated ticket, the conversation and what Risotto tried, and finds what in your setup would have resolved it: an access rule, a runbook template, a tool to finish setting up, or an integration. Insights ranks the findings across the last 90 days, and each one links to the next step or hands it to the Assistant.

### Can Risotto tell us where our knowledge base falls short?

Yes. Risotto Insights reads the last 90 days of tickets and lists questions no article answers, and documents that kept leading to escalations. The Assistant drafts the missing article, and a ✍️ reaction in Slack turns a good answer into a new one.

### Can we ask custom questions about our ticket data?

Yes. Ask Risotto's Assistant in plain English. Code Mode writes the analysis, runs it in a sandbox, and returns the answer with a CSV you can download. You can also pull tickets and both Insights reports into Claude or any MCP client through Risotto's MCP server. [The Assistant](/product/assistant)

### What counts as solved automatically?

A ticket Risotto resolved with no escalation and no person stepping in. Risotto reports it as a share of all solved tickets, apart from the tickets your team resolved.

### How does Risotto calculate time saved?

Risotto takes your average human time to resolve, from escalation to resolution, minus its average AI time to resolve, and multiplies the difference by the number of tickets it solved automatically. Both times come from your own tickets in the period you choose.

## See what your tickets say to automate first

In a pilot, Insights runs on your own tickets.

[Book a demo](/demo)

Source: https://www.tryrisotto.com/product/insights
