Skip to content

Reports & Analysis

Operation statistics

Volume over time: calls per hour, answered vs failed, per callbot and per scenario, with quick ranges (today, 7 days, 30 days). This is the report management asks for; it exports what it shows.

Conversation analysis

The deep report, per scenario (also reachable as Analysis from the scenario itself):

  • Node flow — how many calls entered each node, where they went next, and where they left. The drop-off column is the design feedback loop: a node bleeding callers is a prompt worth rewriting.
  • Node graph — the same data drawn over the scenario's shape.
  • Node trace — individual paths through the flow for sampled calls.
  • NLU evaluation — how the understanding layer performed against the intents the scenario declares.
  • Analysis queue — batch analyses you request over a date range, with their progress.

Activity log

Who did what in the studio, when: scenario saves and publishes, lock takeovers, campaign starts, admin changes. Filterable by user, action, and date — the operational companion to the admin audit trail.

PLXY AI Call Center — built as five separated projects and one shared schema.