Agent setup and testing
By Dan Castiel 12 articles
Customize your Agent from content to interfaceUnderstand every layer of Agent configuration and use a safe sequence to turn brand requirements into tested behavior.Test your Agent in PlaygroundRun realistic, repeatable conversations in the isolated Playground and diagnose failures at the correct source.Shape your Agent guidanceConfigure personality, answer length, language, rules, moderation, and follow-up behavior without duplicating facts.Customize the chat interfaceApply accessible brand styling, customer-facing text, and page-specific entry-point rules, then validate live behavior.Configure tools, surveys, and chat invitationsEnable only tested Agent actions, collect necessary visitor information, and invite shoppers without disrupting the store.Create and run a Batch TestBuild a reusable Batch Test dataset manually, from CSV, or from the catalog, then run it against an Agent.Review and rate Batch Test resultsInterpret Batch Test responses, apply Good, Acceptable, or Poor ratings, add notes, and export evidence.Test how the Agent handles a productUse the product test surface to verify how the Agent presents a real indexed catalog item and its available data.Troubleshoot an incorrect or unsupported answerTrace a weak Agent answer to Knowledge, catalog, guidance, tools, or an unsupported request without adding conflicting fixes.Configure the chat surveySet up a customer-facing survey deliberately, test its timing and wording, and verify the response path before deployment.Configure chat invitationsCreate targeted invitations that open an Agent conversation without overwhelming visitors or duplicating entry points.Configure Agent tools and their dependenciesEnable only Agent actions whose required data source and customer journey are ready Follow a safe workflow, verify the customer-facing result, and isolate common problems.