Tools provided
Tools the agent exposes — your AI client calls them automatically when it needs them.
- run_full — Orchestrated path: cluster your provided accounts, transcripts, tickets and usage into data-backed personas — hard-capped at four — and return the full persona document with coverage warnings.
- analyze — Run pure deterministic clustering over a provided data batch with no model call, returning capped, scored personas plus coverage and injection warnings.
- cluster_personas — Cluster the provided accounts into personas and return each one's identity — id, name, segment and account count — the clustering step on its own.
- extract_jtbd — Extract the Jobs-to-be-Done statements per persona. A theme whose quote negates it is reported as friction rather than a want; a theme with both a want and a push-back is marked contested.
- collect_evidence_quotes — Collect the verbatim evidence quotes backing each persona from your transcripts. Quotes that look like a prompt-injection attempt are flagged in the warnings.
- summarize_usage_patterns — Summarise the top product-usage patterns per persona, including how many accounts contribute to each figure so a single heavy user cannot pass as a shared behaviour.
- map_decision_criteria — Map the decision criteria per persona from the categories of your support tickets.
- list_capabilities — List the agent's static capabilities: the available tools, the provided-data model and the hard four-persona cap. Useful for discovery.
- plan_inputs — Plan the inputs for a tool: returns the clarifying questions, the JSON schema and a ready-to-edit example.
- discover_intent — Understand your goal and co-design the exact input through clarifying questions before running.
- open_form — Opens the guided form to collect the segment and the number of personas. The entry point for running the agent on your own data.
- run_form — Internal: invoked by the guided form when the user submits. Not for direct use.
