Braid

Analytics

What resolved, how fast, and what it could not answer

Four numbers that matter and the reasons behind them — counted across every channel together, in your workspace's own timezone.

  • Conversations handled

    The volume over 7, 30 or 90 days, with the change against the period before it.

  • Auto-resolved by AI

    How many finished without a person. The single number most teams are actually buying.

  • Median first response

    A median rather than a mean, because response times are long-tailed enough that the two say genuinely different things.

  • Handoff rate, with reasons

    Not only how often it hands over, but why — asked for a person, frustrated, or not in your sources.

  • Topics

    What customers are actually asking about, and which topics the agent resolves on its own.

  • Export it

    CSV of the period you are looking at, for the spreadsheet someone will inevitably want.

Watch the reasons, not just the rate

A rising 'not in your sources' is a knowledge problem. A rising 'frustrated' is a different one entirely.