This custom law firm chat bot is a production-grade, fully client-side conversational intake engine built in hand-written vanilla JavaScript, with zero framework dependencies. It was built as an IIFE wrapping a class-based controller, and singleton service modules, with immutable configuration, making it lean, fast, and self-contained.

At its core, the chat bot is a finite-state conversation engine that guides users through case-type selection, step-by-step contact capture, SMS consent, and a structured incident interview which covers injuries, treatment, police reports, and location.

On top of Compromise.js sits a custom natural-language layer that handles fuzzy interpretation, name and city extraction, accident-type and injury classification, severity and sentiment scoring, and automatic triage of out-of-scope cases. Every input is hardened and validated: area-code-aware number checks, approved-and disposable-domain email filtering, and fuzzy matching of cities against the firm’s licensed service areas. XSS is neutralized through textContent escaping and template cloning, and per-state retry guards keep the flow resilient.

The system orchestrates real multi-channel automation featuring server-mediated Twilio SMS notifications and timed follow-up drips, CRM integration across Clio, MyCase, Lead Docket and more, and inline Calendly or Cal.com booking. Rounding it out are a rate limiter, CSRF protection, session persistence, conversation history, and offline-aware network handling.