AI voice agent can automatically route calls to the right person or team based on what the caller says. To set it up, you need to configure a flow based on pre-defined keywords.
What are keywords?
Keywords are specific words or phrases the AI agent listens for during a call. When the agent hears a match, it triggers an action like transferring the call to:
A specific user
Another AI agent
How to choose keywords
Keywords trigger when your AI agent should route a call to a specific person or team. The goal is to match how real callers actually speak, not just internal labels.
1. Start with common caller language
Think about how customers naturally describe their needs, not how your team names departments.
Instead of just “billing”, include “invoice”, “payment”, “charge”, “refund”, “bill”.
Instead of just “support”, include “problem”, “issue”, “not working”, “help”, “error”.
Tip: Review past calls, emails, or tickets to find real phrases customers use.
2. Include variations and synonyms
People say the same thing in different ways, so include multiple variations.
For example, keyword for sale team can be:
“pricing”
“cost”
“plans”
“quote”
“how much”
3. Use intent-based keywords, not just topics
Most of the time, callers clearly state what they want to do, not just the topic. Include action-oriented phrases to increase the chances of an accurate match.
For example:
“book appointment”
“schedule a call”
“talk to someone”
“speak to sales”
4. Group keywords by destination
Each transfer destination should have its own clear set of keywords.
Example:
Sales team: “pricing”, “quote”, “plans”, “buy”, “demo”
Support team: “issue”, “problem”, “not working”, “error”
Billing team: “invoice”, “payment”, “refund”, “charge”
5. Avoid overly broad keywords
Generic words can cause incorrect transfers. Avoid using keywords like “help”, “question”, “information”. These can apply to almost any situation.