Introducing the new Quality Monitor: every alert investigated by Enterpret Agent

Today we're launching the new Quality Monitor, powered by Enterpret Agent. The agent now investigates every possible quality issue before it reaches you. In our benchmark, that cut false positives by 95% and produced 40% more accurate alerts. When an issue is real, it can trigger work in your systems, like a Linear or Jira ticket or a PagerDuty incident, with the evidence attached. It's live today for all Enterpret customers.

Quality Monitor watches your customer feedback for things that break. That might be a jump in checkout complaints after a release, or a bug that shows up in support tickets and app reviews in the same afternoon. When one appears, it alerts your team.

It started as a spike detector. It tracked how often each topic came up, flagged the ones that jumped, and wrote the alert from summaries of the feedback. A spike shows you where to look. But someone still had to confirm the issue was real, find what changed, count who was affected, and write it up for the team that owns the area. Enterpret Agent now does that work.

95% fewer false positives

When Quality Monitor detects a spike, Enterpret Agent opens a session on it. The agent:

• reads the feedback records behind the spike
• measures how many customers are affected
• checks your knowledge base to tell expected behavior from breakage
• looks up recent releases when it needs to

If a candidate doesn't hold up, it isn't sent.

What it filters out is the noise teams told us about: an outage you already resolved, one customer repeating the same complaint, a viral thread, a feature working as designed. In our benchmark, false positives fell by 95%.

40% more accurate investigations

The same benchmark produced 40% more accurate alerts. Most of the gain came from issues that were real but had been described wrong. Because the agent reads the records themselves, it scopes each issue to the customers it actually affects and names the cause.

Every alert carries a chart, a root cause, and the records behind it. Each one is also a session you can open, to see what the agent read and keep asking questions. And because Quality Monitor runs on Enterpret Agent, improvements to the agent show up in your alerts too.

Alerts that trigger work in your systems

Quality Monitor can now trigger work, like any other automation, and you choose what a real issue sets in motion:

• A Linear or Jira ticket for the owning team, filed with the root cause, the customers affected, and their own words. If a matching ticket is already open, the evidence is added to it instead.
• A PagerDuty incident for regressions that meet your on-call criteria.
• A Slack or email alert with a trend chart, as before.

You set it up once from a template, and it runs on every issue that matches your filters.

Why we built it

An alert asks someone to stop what they're doing, and we wanted every Quality Monitor alert to be worth that.

Quality Monitor also used to end at a message, and customers were already building the next step themselves. One team copies our Slack alerts into Jira by hand. Another took an alert about a feature getting stuck and drafted safer guidance for its support team. It also followed up with three customers whose tickets had already closed, all within two hours. The detection was useful both times, and people did the rest.

Quality Monitor is the second piece of the customer loop we're building on Enterpret Agent. Escalation Shield triggers work for one customer who's about to escalate. Quality Monitor triggers it for an issue reaching many. Next, the agent will measure an issue again after the fix ships, to confirm it worked.

If you already use Quality Monitor

Your monitors, filters, and Slack or email channels stay as they are. Alerts now come from the investigation, and each monitor appears on the Automations page, where you can add an action.

The investigation behind each alert runs on AI models, so a sent alert uses AI credits like any other automation. Investigations that end without an alert aren't charged. Visit the help center for more details.

Get started

The new Quality Monitor is live today for all Enterpret customers. Open a monitor on the Automations page and pick a template for what a real issue should trigger, or tell Enterpret Agent what should happen when it finds one.

Not using Enterpret yet? Request a demo and we'll show you the whole loop, from detection to investigation to the issue landing in your systems.