Measure

Calls and analytics

Read real calls turn by turn, watch trends across them, and use both to make the agent more reliable.

SurfaceWhat it shows
ConversationsOne row per call, opening into the turn-by-turn conversation, the recording and the per-call checks.
DashboardsTrends across calls on a dashboard you arrange.

Start with a bad call under Conversations for why, and Dashboards for how often.

When a call appears

Calls land within a minute or two of finishing. The recording has to upload and the analyser has to run, so a recent call may read as still processing.

Each call stores the turn-by-turn transcript, tool calls and analysis (summary, sentiment, outcome, duration). When recording is enabled, it stores the audio too. Sensitive details are redacted before storage. See Privacy.

The call list

The Conversations page lists calls, newest first. Open it from an agent and it's filtered to that agent. Each row carries the headline signals:

  • Contact (name or caller number), with a Test tag on calls your team placed from the app.
  • Agent, Summary, Outcome (completed, abandoned, transferred or failed, or a silent call when the caller never spoke), Date and Duration.
  • Checks, a tally of the pass-or-fail checks the call passed.
  • Reviewed, once someone has marked the call reviewed.

The Needs you tab narrows the list to unreviewed calls that need a look, with a Why column giving the reason.

Filtering, saved views and export

The toolbar gives you Search, a views menu, an Agent filter and Filters: date range, review status, outcome, caller mood, topic, channel, source and silent calls.

Save current view in the views menu stores the current filters so you can return to them, and Export this view as CSV downloads the list. The list defaults to Customer conversations, without silent calls. The Source filter switches to test conversations, scenario runs, or everything except scenario runs.

Reading a call turn by turn

Click the Checks tally to open that call's checks, anywhere else on the row to open the call in a side panel. The panel steps to the previous or next call, and its arrow opens the full page.

The header shows the caller, the topic, the outcome, channel, time and conversation length. Below it sit the call's Summary and Result, then the recording. On a wide screen the conversation sits on the left, with the Checks, Activity, Performance and Details tabs on the right.

Tool calls in the transcript

When the agent called a tool, a chip sits above the agent's reply, coloured green when the tool succeeded and red when it failed. Tap it to inspect the tool's arguments and what it returned.

This is usually enough to answer "why did it say that". The agent's prompt, its tools and what the caller said are all on the page.

The recording

For voice calls with recording enabled, an audio player sits above the conversation with a waveform and separate lanes for agent and caller. The player's clock shows the recording's own length, which can differ from the conversation length in the header.

If a call just ended, the recording shows "Recording processing" until the upload finishes. Text chats have no recording. Recording is governed by your Privacy settings.

Recording disclosure

Whether callers are told the call is recorded is part of your privacy settings, not something you set per call. Set the spoken disclosure in Privacy.

Per-call scores

The Summary card holds the call summary, the caller's mood, any failure reason and what the agent did. The Result card shows Task success, Hallucination, Caller satisfaction (CSAT) and Median reply time when they were scored, then how many checks passed and which failed.

The checks tab

The Checks tab lists every check on the call. Filter it by All, Failed, Passed or Measurements. Checks are grouped into:

  • Decides the result, pass-or-fail checks that set the call's result.
  • Doesn't decide the result, pass-or-fail checks kept for review.
  • Measurements, numbers such as reply time.
  • Scores, ratings such as caller satisfaction.

Scoring on demand

If a call was never scored, Run checks scores it now. Run again rescores a call that was already scored. Flag a verdict you think is wrong with the thumb on its row, and note what it should have been. How calls are scored is covered in Call checks.

Using call records in practice

  • Find a failure. Open a bad call and read it back: a misread intent, a tool that returned an error (the red chip), a clumsy hand-off. The recording often makes a tone or timing problem obvious that the text alone hides.
  • Spot a pattern. Filter to failed or transferred calls and read a handful. The same stumble across several calls is a prompt or tool fix waiting to happen.
  • Confirm a fix. After you publish a change, watch new calls and check the same failure doesn't come back.

The improvement loop

  1. Find the weakest behaviour in a bad call or a sagging trend on the Overview.
  2. Write an eval for it so the failure is pinned.
  3. Fix the cause: tighten the prompt or a tool description (tools first, prompt last), then publish.
  4. Watch the new calls. Confirm the fix holds and the trend moves.

On this page