AI replies shaped by your content and settings
Connect a managed or customer-owned OpenAI key, set system prompts, choose model settings, and feed answers from your knowledge base.
Dchat is built around one practical support loop: answer routine questions with AI, hand off the exceptions to humans, and manage everything from one workspace.
Connect a managed or customer-owned OpenAI key, set system prompts, choose model settings, and feed answers from your knowledge base.
When AI hits a boundary, the conversation moves to a human teammate with the transcript intact so visitors do not have to repeat themselves.
Start with the hosted stack for fast rollout, then move to Zchat.com if you need a self-hosted path later.
That matters because the best support setups are not just about answering faster. They also make it easier for the team to adjust prompts, content, routing, and handoff behavior without starting over.
AI handles the routine layer with content-backed replies and clear boundaries.
Humans step in with the transcript intact, so the conversation never feels reset.
Knowledge, branding, routing, and reports stay in one place instead of spreading across separate tools.
What the support workspace includes
Live inbox
Real-time updates keep AI-owned, waiting, and resolved conversations visible in one place.
Canned responses
Save repeatable replies so your team can move faster without losing consistency.
Routing rules
Keyword and handoff rules keep the right threads moving to the right queue.
Knowledge base
Feed AI answers from manual articles, imported URLs, or site crawl content.
Offline capture
Collect messages when nobody is online so visitors still leave a path back to the team.
Contacts
Visitor identity details flow into contacts so conversations can become ongoing relationships.
Reports
Track conversation volume, response timing, ratings, and AI usage from the same workspace.
Workspace settings
Tune widget branding, AI settings, routing behavior, retention, and billing in one admin surface.