> For the complete documentation index, see [llms.txt](https://converso.gitbook.io/product-documentation/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://converso.gitbook.io/product-documentation/ai-agents/agent-store/customer-support-agent.md).

# Customer Support Agent

AI customer support agent answers customers' queries based on its knowledge of your business and its products.

### Who's it for

SaaS businesses.

### The challenge

Customers and evaluating visitors expect instant, accurate answers 24/7 - on your help pages, docs, and post-sale channels. They want product clarity without a sales pitch, and a first-line support team can't cover that volume around the clock.

### How it works

The Customer Support Agent answers questions based only on your company's own knowledge - no "hallucinating" or made-up information. It helps existing users understand and use your product, and answers evaluators' questions factually, without a sales motion. It never qualifies visitors, captures leads for nurturing, or pushes demos.

Your Customer Support Agent can be deployed over webchat and WhatsApp (other channels coming). Once published, all new incoming messages on connected channels are assigned automatically to the agent, which responds directly to your customer.

### Where to deploy

Help centres, documentation sites, and post-sale pages - anywhere people come for answers, not pitches. For pages where conversion is the goal, deploy the [Sales Agent](https://converso.gitbook.io/product-documentation/ai-agents/agent-store/sales-agent) instead.

### Knowledge base

Give it your product documentation, feature facts, FAQs, and how-to guides. Keep sales collateral out - a factual knowledge base is what keeps its answers neutral and trustworthy.

### Key capabilities

**Product & pricing questions** - The agent draws on the documents, URLs, and FAQs in your Knowledge Base to answer detailed questions about what you offer and how much it costs - accurately, and grounded in your approved product facts.

**Feature explanations** - When customers want to understand how something works or what's included, the agent gives clear step-by-step guidance, tailored to the tone and technical level you've configured.

**Issue triage** - The agent acknowledges problems, walks customers through your documented troubleshooting steps, and hands unresolved issues to your team promptly rather than running extended diagnostics.

**Sales enquiry detection** - When a visitor shows clear buying intent - a demo or trial request, custom pricing, or asking who to talk to about buying - the agent answers what it factually can, then offers to hand the conversation to your team, flagged as a sales enquiry with the context summarised. It detects and hands over; it never runs a sales conversation.\
\
In cases where there is no human available to take over the chat, a contact form is sent after 120 seconds to the customer to collect their name and contact details.

### AI-to-human handoff

If the question is too complex for the AI agent to answer, it's trained to recognise when its answer is not sufficient to keep the customer happy, and it will offer to hand over the chat to a human agent. Alternatively, the customer can just ask to chat with a human agent directly. It's also possible to assign the chat back to the AI again, if required.

All AI agent chats with customers are streamed in real time in the workspace team inbox, whether assigned to the AI Agent or a human agent, and a human agent can jump in at any time to unassign the chat from the AI agent and take over the conversation with the customer, or assign the chat to another team member. 

### Best suited for

SMB SaaS and tech services businesses with a high volume of support and product questions on help pages and post-sale channels. Particularly effective where customers need product clarity in a setting where a sales approach would feel wrong.