A first look, an ongoing check, or a deeper human perspective.
Choose the level of evaluation your support agent needs.
See how your agent handles a small set of challenging support scenarios.
AI evaluation · Human review available separately
Make evaluation a regular part of running and improving your support agent.
Proposed introductory price · No payment collected in this preview
Independent human judgments, scoped around your needs.
Independent verdicts, preferences, and explanations on selected test results.
Custom scopeHow it worksHuman assessment of real support interactions, with an agreed sampling and review process.
Custom scopeHow it worksA minute measures the time spent running test conversations with your support agent. It is not a number of questions or a human-review minute. Detailed metering and overage rules will be confirmed before a paid subscription starts.
The free evaluation and Continuous plan use AI evaluation. Independent human review is a separate service, scoped around the conversations, criteria, and volume you need.
The starting point is your website chat URL and the context needed to test it. The automated tester uses the customer-facing chat. If your agent needs a login or restricted access, the test setup needs that access too.
A conversation shows what the agent said. Confirming a refund, booking, or other external action requires supporting evidence or access to the relevant system. Reports distinguish the claim from a verified outcome.
Live-conversation review means people assess selected production interactions. Real-time approval or interception is not included in the offer described here; the access and review cadence are agreed separately.