Skip to content
SUPERWISE Chat SUPERWISE Chat SUPERWISE Chat

Launch Scenarios

These are the core use cases SUPERWISE Chat is built to handle at launch. Each scenario has been exercised end-to-end in the platform’s integration test suite (tests/scenarios/run-feature-scenarios.ts).

  • Persona-guided decision support — Apply a named behavioral style (Signal, Clarity, or Depth persona) to shape how the assistant responds: concise executive summaries, plain-language explanations, or comprehensive deep-dives.

  • Business assistant Q&A — Ask multi-turn questions through a configured Business Assistant; the assistant maintains context across turns and can switch to a structured advisory mode (e.g., Analyst persona) mid-conversation.

  • Project-scoped conversations — Create or join a Project (e.g., Engineering Hub, Strategy & Leadership) to scope conversations, knowledge, and assistant defaults to a specific team or initiative.

  • Web search — Send a query that requires up-to-date information; the platform fetches and distills live web results before responding. Search is tier-controlled: QUICK and THINKING run one search pass; RESEARCH and THOUGHT_PARTNER run multiple.

  • Multi-turn context retention — Continue a conversation across multiple messages with the assistant correctly referencing earlier turns, including back-references to statements made several messages ago.

  • Guardrail enforcement — Submit a borderline or sensitive request; the platform evaluates it against configured guardrail rules, responds safely, and records the decision in the governance trail.

  • Real-time streaming events — Receive live progress signals during a response: thinking steps, token spend increments, and guardrail results arrive via Server-Sent Events (SSE) as the pipeline runs, not after it finishes.

  • Governance tracing — After any conversation, inspect the per-message execution trace in the Governance Panel to see which pipeline phases ran, how long each took, and what guardrail decisions were made.

  • Knowledge retrieval — Attach a knowledge collection to a channel or project; the platform retrieves relevant chunks from indexed documents during the Gather phase and cites them in the response.

  • Knowledge + persona framing — Combine a knowledge collection with a persona overlay so retrieved content is presented in the persona’s style (e.g., Analyst persona formats retrieved engineering docs as a structured technical review).

  • Full feature stack — Use all major capabilities together: a configured assistant, a writer persona, and a project scope active simultaneously, with governance tracing and knowledge retrieval running in the same conversation.