01
Conversation ends
The visitor closes the chat, hangs up the call or the Telegram thread is closed. Nothing is asked of them.
How it works
What happens after the visitor closes the tab
- 02
A reviewer reads it
A reviewer goes through the transcript — including the parts no human would scroll back to.
- 03
Fields get filled
Summary, outcome, sentiment, score, action items, knowledge gaps and tags — or your own custom fields — are filled from the transcript.
- 04
Actions can run
A reviewer you set up can add a lead or ticket to the Datastore, or run a follow-up in a connected app.
Capabilities
What a review captures
A rich record you can skim, search and act on — written for someone who wasn't there.
Summary
What the visitor wanted, what the agent did, and how it ended — in a few readable sentences.
Outcome and score
Whether the conversation completed, stalled, was cancelled, needs follow-up or was escalated — plus a 0–100 score for how well it met the visitor's goal.
Sentiment
How the visitor felt across the whole conversation, from very negative to very positive — not just how politely it ended.
Next steps
Action items with a low, medium or high priority — plus an owner and due date when the conversation names them.
Knowledge gaps
The questions your agent couldn't answer from its knowledge — a content to-do list written by your own customers.
Tags
Short labels so themes show up in your analytics without anyone tagging by hand.
In practice
Two reviewers, one job
Keep the built-in reviewer, or write your own with custom fields and follow-up actions.
The default reviewer
Zero setup. Every conversation comes back with summary, outcome, sentiment, score, key points, action items, knowledge gaps and tags.
A reviewer you define
Write instructions in plain language, define your own fields, and give the reviewer the Datastore, knowledge search or connected-app actions.
Save to Datastore
Add leads, tickets, feedback or any form you design, so the outcome lands where your team already works.
Never blocks the visitor
The review and its actions run after the conversation closes. A slow follow-up never freezes a live call.
FAQ
Questions
When does it run?
Automatically, as soon as a chat, call or Telegram thread ends — or after a period of inactivity if the visitor just walks away. The visitor never waits on it; the review happens in the background.
Do I have to set anything up?
No. A built-in reviewer runs from day one with sensible defaults. Replace it with your own when you want custom fields.
Can I change what gets extracted?
Yes. Write your own reviewer, describe what to look for in plain English, and define the exact fields you want back. If it returns nothing usable, the built-in reviewer fills in instead.
Can it do more than fill a record?
Yes. Give your reviewer the Datastore or connected-app actions you choose and it can add a lead or ticket and run follow-ups — all after the visitor is gone.
Does it work for voice calls?
Yes. Browser voice calls are transcribed as they happen, so a call is reviewed exactly the same way a chat is.
Will it invent things that weren't said?
The reviewer is held to the transcript. If the conversation gives it no basis for a field, it leaves it out rather than guessing.
Can different agents review differently?
Yes. The reviewer is chosen per agent, so support and sales can extract completely different things from their conversations.
Where do I see the results?
Each conversation shows its review on the conversation page. Scores, sentiment, tags and knowledge-gap counts also roll up into your analytics dashboard.
Free plan · paid plans free for 7 days
Build your first customer agent.
No credit card. Questions? Write to support@embage.com.