> ## Documentation Index
> Fetch the complete documentation index at: https://docs.thunderphone.com/llms.txt
> Use this file to discover all available pages before exploring further.

# AI reports

> Ask questions in plain English and get a report built from your actual call transcripts — funnels, drop-offs, and answers.

The **Reports** page (`/dashboard/reports`) under **Quality & Testing**
turns call transcripts into analysis. Pick the agents and date range,
optionally ask specific questions, and the AI reads the calls in scope
and reports back — call funnel, where callers drop off, recurring
patterns, and direct answers to whatever you asked.

## Generate a report

<Steps>
  <Step title="Scope it">
    Choose one or more **agents** and a **date range** — last 7/30/90
    days, all time, or custom dates. Selecting several agents produces a
    comparison report.
  </Step>

  <Step title="Ask your questions (optional)">
    Add anything specific you want answered, up to 2,000 characters —
    e.g. *"How often do callers ask about pricing? Did the July script
    change help?"* Suggestion chips cover common asks like "Where do
    callers drop off, and why?" and "Compare my agents against each
    other."
  </Step>

  <Step title="Exclude noise (optional)">
    Under **Exclude numbers from analysis**, toggle off labeled numbers
    (your own team's test lines, for instance) or paste numbers to
    exclude.
  </Step>

  <Step title="Click Generate report">
    Generation takes a few minutes for large call volumes. The report
    analyzes up to the **200 most recent calls** in scope.
  </Step>
</Steps>

## Quota

Reports are limited to **3 per organization per calendar month**. The
composer shows how many you have left and when the quota resets (the
1st of the next month). Spend them on well-scoped questions — a report
with a specific ask beats a generic one.

## Reading and reusing reports

Generated reports are saved on the page. Open one to see the call
funnel, metrics, distributions, and the answers to your questions. From
a report you can:

* **Re-run** it on a new date range with the same agents, questions,
  and exclusions — the quickest way to track a change over time.
* **Chat in your AI instead** — export the underlying data and continue
  the analysis in ChatGPT, Claude, or Gemini. Exports don't count
  against your report quota.

## Next steps

<CardGroup cols={2}>
  <Card title="Observability" icon="chart-line" href="/guides/observability">
    Live operational metrics — the numbers behind the narrative.
  </Card>

  <Card title="Issues" icon="triangle-exclamation" href="/guides/issues">
    Triage the recurring problems reports surface.
  </Card>
</CardGroup>