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

# Analyzing agent behavior

> How The Context Company turns captured production data into a workflow for discovering, understanding, and quantifying agent failures.

The Context Company captures the [full execution](/what-tcc-captures) behind every agent interaction. This section covers the four surfaces you use to turn that data into decisions.

## The four questions

Each analysis feature answers a different question about production behavior.

<CardGroup cols={2}>
  <Card title="Patterns" icon="magnifying-glass" href="/analyze/patterns">
    **Is behavior we already care about happening?** Reusable classifiers for known behaviors: frustration, confusion, task failure, or any custom signal you define.
  </Card>

  <Card title="Failures" icon="triangle-exclamation" href="/analyze/failures">
    **What recurring problems keep happening?** Errors from runs and tool calls, grouped and ranked by frequency and recency.
  </Card>

  <Card title="Insight Search" icon="comment-question" href="/analyze/insight-search">
    **Why is this happening?** Natural-language investigation across runs, sessions, tools, models, costs, and feedback.
  </Card>

  <Card title="Recaps" icon="chart-line" href="/analyze/recaps">
    **What changed that I should pay attention to?** AI-generated daily, weekly, or monthly reports with linked evidence.
  </Card>
</CardGroup>

## Dimensions you can slice by

Every analysis works across the same first-class dimensions, because they are stored as native columns rather than free-form metadata.

<CardGroup cols={2}>
  <Card title="Conversations and sessions" icon="messages" href="/analyze/conversations-and-sessions">
    Runs grouped into full user conversations.
  </Card>

  <Card title="Users and organizations" icon="users-viewfinder" href="/analyze/users-and-organizations">
    End users and the customer accounts they belong to.
  </Card>

  <Card title="Tools and models" icon="wrench" href="/analyze/tools-and-models">
    Which tools and models were involved, with cost, latency, and error breakdowns.
  </Card>

  <Card title="Feedback" icon="thumbs-up" href="/analyze/feedback">
    Thumbs up / down and text feedback per run.
  </Card>
</CardGroup>

## A typical workflow

<Steps>
  <Step title="Something surfaces">
    A Recap flags a spike, a pattern crosses its alert threshold in Slack, or you notice a change on the overview dashboard.
  </Step>

  <Step title="Group the affected runs">
    Open the pattern or failure. Both list the runs (and sessions) that match, ranked by recency and frequency, with links into each trace.
  </Step>

  <Step title="Investigate root cause">
    Open a representative [trace](/investigate/traces). Inspect the model steps, tool calls, arguments, results, and errors that led to the outcome.
  </Step>

  <Step title="Quantify impact">
    Filter by [user, organization](/analyze/users-and-organizations), [tool, or model](/analyze/tools-and-models) to see how many customers are affected and where the problem concentrates.
  </Step>

  <Step title="Ship the fix and monitor">
    Change the prompt, tool, or workflow. Create a [custom pattern](/analyze/patterns#custom-patterns) to track the behavior going forward, and let Recaps and Slack alerts flag regressions.
  </Step>
</Steps>

## Related

<CardGroup cols={2}>
  <Card title="Traces" icon="diagram-project" href="/investigate/traces">
    The primary evidence behind every analysis result.
  </Card>

  <Card title="Silent tool failure walkthrough" icon="bug" href="/investigate/tutorial-silent-tool-failure">
    Discovery → trace → root cause → monitoring, end to end.
  </Card>
</CardGroup>