Upgrading your support agent

Your support agent is now ready to answer user queries. However, you can combine its ability to answer questions with other features to build an even more powerful agent. Here's some suggestions:

‍

Dynamically route the user based on intent

By enabling the Listen for other intents option on your Capture step, you can automatically route a user's query to another workflow. For example, if you're building a support agent for an online retailer, if the user says "I want to return my order", you could trigger a Return Order intent and route them to a specialized workflow. You can learn more about routing users between workflows based on intents in the Voiceflow Basics course.

‍

Gracefully handle follow-up questions

Sometimes, users might ask follow-up questions. For example, they could first ask "what's your returns policy?", then after the agent answers, they could ask "does that include food products?". Currently, our agent can't handle this situation, as it only optimizes a single query. Check out this template to learn how to handle follow-up questions.

‍

Connect to third-party services

While AI agents are powerful, they won't always be able to solve every problem alone. In some situations - such as if the agent repeatedly can't find an answer to a user's question - you might want to route them to a human via a service such as Zendesk or Freshdesk.

‍

To add this functionality to your agent, you could use variables to count how many times the user received an error, and after a certain amount, use a function to create a support ticket. We have pre-made functions for Zendesk and Freshdesk, but you can connect your agent to any ticketing platform that has an API.

‍

For a complex example of this in action, check out our Fintech ticketing system that connects to Google Calendar, Zendesk and Airtable via n8n.