How it works

A short rollout path, not a giant support migration.

Dchat is designed to get teams from widget embed to a working AI + human support loop without turning setup into a project of its own.

1

Embed the widget

Add the loader to your site, set the widget token, and publish the launcher where visitors already expect support to live.

<script src="loader.js" data-widget-token="...">

2

Connect AI and content

Choose Dchat-managed AI or bring your own OpenAI key, then add knowledge base articles manually or through URL import and site crawl.

AI settings Knowledge base Site crawl

3

Tune the team workflow

Adjust routing, branding, proactive chat, canned responses, reporting, and seat settings from the same workspace after the initial embed.

AI Settings Routing Widget Contacts Reports Billing

A small team planning together around laptops in an office — **A rollout teams can actually finish** The goal is to get from embed to a working support loop quickly, then improve from live traffic instead of delaying everything up front.

Why the rollout stays short

The setup path is practical because it follows the real order teams work in.

Put the widget live, give the assistant useful content, then tune the human workflow once real conversations start showing you where the exceptions are.

Publish the entry point

Get the widget in front of visitors where they already expect support to live.

Ground the answers

Knowledge, prompts, and AI settings make the first reply more useful before the team ever joins.

Refine from evidence

Routing, reports, and handoff patterns improve once the real conversation mix becomes visible.

Platform Features