Set up a voice agent your team can trust.
Kawl Now guides cover the practical side of running an AI voice agent on a SaaS website: adding it to your site, building the knowledge base it answers from, writing its instructions, testing it before customers call and managing it from your coding tools over MCP. Each one follows how the product works today, including its limits.
Reviewed · By Tenhaw
- Install guide
How to add an AI voice agent to your website
Add an AI voice agent to your website with one script tag. Approve its knowledge, set it up, allow your site and test, with 50 free conversations to start.
Read the guide - Practical guide
Build an AI knowledge base for customer support
Build an AI knowledge base for customer support: choose the questions, write answers your agent can find, review GitHub docs and test before customers call.
Read the guide - Voice agent testing
Voice agent testing: how to test before launch
A voice agent testing plan for your website: real customer questions, browser and microphone checks, and how to read Kawl Now transcripts to fix weak answers.
Read the guide - MCP setup guide
Manage a voice agent from a coding agent over MCP
Connect your coding agent to the Kawl Now MCP server to set up your website voice agent: add knowledge for review, change settings and get the embed code.
Read the guide - Guide
Turn your FAQ page into an AI voice agent
Turn your FAQ page into an AI voice agent that works like a spoken FAQ chatbot: rewrite answers to be heard, group them into sources, approve and test them.
Read the guide - Guide
Knowledge base article template for AI support
A knowledge base article template for AI support: one question per article, the answer first, customer words and conditions, plus a filled-in example to copy.
Read the guide - Guide
Write voice agent instructions and greetings
How to write voice agent instructions and greetings, with system prompt examples for a support agent and a sales agent to copy, adapt and test.
Read the guide - Guide
Run separate AI support and sales agents
Run separate AI support and sales agents on your website: split the knowledge, give each its own instructions and pages, then test and review each one.
Read the guide - Guide
Voice agent fallback responses: when it can’t answer
Write fallback responses for a voice agent that cannot answer: say so plainly, give the right contact route and note feedback, with example wording to test.
Read the guide - Guide
Localhost testing and staging for a website voice agent
Localhost and staging testing for a website voice agent: allow each exact origin, use HTTPS for phones and keep production off the list until launch day.
Read the guide - Guide
Voice agent launch checklist for your website
A voice agent launch checklist for your website: what to check before launch, on launch day and in the first week, plus how to switch the agent off fast.
Read the guide - Guide
Conversation analytics: review your voice agent’s calls
Conversation analytics for a voice agent: read the transcripts and automatic insights each week, label each outcome and fix the docs behind weak answers.
Read the guide - Guide
Customer service KPIs for an AI voice agent
Which customer service KPIs to track for an AI voice agent, how to define each one and how to work them out from your conversation export and transcripts.
Read the guide - Guide
Docs as code for AI support: keep answers current
Use docs as code to keep AI support answers current: import GitHub docs, re-sync after each merge, approve the new text and handle renamed or deleted files.
Read the guide - Guide
API key security: keep workspace tokens safe
API key security for your voice agent: store each workspace token as a secret, give every tool its own, rotate on a schedule and know what to do after a leak.
Read the guide