Welcome, mufty.851.
This is the staff dashboard for the AI intake chatbot. More screens (leads, conversations) arrive in upcoming milestones.
What you can do now
Structured, scored leads the AI pulled from chatbot conversations — read the transcript, set the status, and work your pipeline.
Open Leads →Read every real chat visitors had with the bot — set a review status, flag, tag, assign, and leave internal notes for the team.
Open Conversations →See live metrics from real chatbot conversations — engagement, KB-grounding rate, guardrail flags, and outcomes.
Open Analytics →Add, edit, and organize the FAQs the chatbot answers from — changes are indexed automatically and you can preview retrieval.
Open Knowledge Base →View, edit, version, publish, and roll back the prompts that shape the chatbot's system prompt — with a live preview of what's assembled today.
Open Prompts →See the configured chat + embeddings providers and run a live health check.
Open LLM Settings →Paste your Terms of Use and Privacy Policy page URLs — the chat widget's consent screen picks them up automatically.
Open Widget Settings →Run the pre-built adversarial corpus (PHI, jailbreaks, implied representation, Louisiana-specific traps) against the real chat pipeline and see a category-by-category pass rate.
Open Safety Eval →