3CX AI agent: resolving routine requests using your own knowledge
Opening hours, documents, simple troubleshooting: the 3CX AI agent answers recurring questions from your own content. How it works and how to prepare.

Opening hours, documents to provide, basic troubleshooting: a large share of the calls an SME receives are about recurring questions. The 3CX customer service AI agent answers these requests using the content you give it, whenever you choose to route calls to it. How does it work, and how should you prepare it?
Beyond the auto attendant
An AI receptionist understands the caller in natural language and directs the call to the right destination. The customer service AI agent goes further: it listens to the request, searches for the information in the knowledge sources assigned to it and tries to provide a useful answer before deciding whether human intervention is needed.
3CX positions it for level 0 and level 1 support tasks: frequently asked questions, practical information and the first steps of troubleshooting.
A guided conversation, step by step
As Ruth Elizabeth Abbott, COO of 3CX, points out, the agent asks targeted follow-up questions and guides the caller step by step through basic troubleshooting. The goal: to resolve routine requests before a person needs to step in. In practice, the agent:
- greets the caller and identifies their request;
- asks one question at a time to clarify the need;
- searches for the answer in the configured knowledge sources;
- gives a short answer or one to three practical steps;
- transfers the call when the request requires human support.
Ruth Elizabeth Abbott also stresses that not every request is meant to be handled by AI. The agent is designed to recognise its limits and hand over at the right time.
Your knowledge, your control
The agent answers using information you control: FAQs, internal policies, product guides, support documentation, troubleshooting procedures or internal notes. When the option is enabled, it can also search the web. This is essential for an SME: the answers given to your customers are based on your own content, not on generalities.
Targeted, well-described sources
The quality of the answers depends directly on the quality of the sources. 3CX recommends dedicating each knowledge source to a single topic, with a clear description, and keeping unrelated information separate. Data that changes constantly, such as real-time stock levels, is better served by an API or a database than by a static file.
Up-to-date content
A modified source must be refreshed in the 3CX console. The choice of AI provider also matters: if you change provider, the links to knowledge sources are reset and must be reconfigured. It is therefore best to settle this choice from the outset.
Where to place the agent in your call flows
Customer service is not the only area of application. 3CX mentions schools, medical centres and shops, among others. In an SME, the agent can, for example:
- provide contact or support information;
- tell the caller which documents to prepare before speaking to a member of staff;
- guide the caller through simple troubleshooting steps;
- collect details of a problem before routing the call;
- direct the caller to the right human destination when necessary.
Before configuring anything, five questions shape the project: how do calls reach the agent? Which questions should it handle? Which sources should it use? Where should calls go when human help is needed? Which topics require a specific destination?
Technical prerequisites
The customer service AI agent requires the 3CX AI edition, version V20 Update 10 or later, and a supported real-time API provider.
Preparing the project with MLS Communications
A successful AI agent is above all a matter of organisation: identifying recurring requests, structuring content, defining transfer rules and testing with real cases. We support you at every step, from analysing your needs to going live, and then with continuous optimisation. Our learning-by-doing approach helps your teams understand how the agent works and develop its knowledge sources independently.
Would you like to hand your routine requests over to an AI agent? Contact us to identify suitable requests and prepare the knowledge sources for your future 3CX agent.


