In this tutorial, you'll learn the basics of training and optimize Fin to align with your company's unique voice, processes, and support policies.
Step 1: Fin can be trained to reflect your brand's unique voice and processes. For instance, if your company refers to customer plans as subscriptions, you can guide FIN to use the same terminology.
Step 2: Notice the four types of guidance you can give Fin. The first one, communication style, allows you to dictate the terminology and tone of voice Fin should use. This is where you'll guide how Fin talks to your customers.
Step 3: After adding your guidance, hit the Optimize button. This allows our AI writing assistant to check for issues and suggest improvements.
Step 4: Test the refined version in the preview. If it's satisfactory, enable it and set the guidance live ✔️
Step 5: You can also add data attributes in your guidance to create contextual behavior. For instance, you might instruct Fin to use Hiberno English if the customer is in Ireland. To do this, hit the icon that allows Fin to check an attribute, in this case, the customer's country.
Step 6: Again, hit Optimize and use the preview to see the changes in action. If it's satisfactory, hit Enable.
Step 7: To create a more personalized customer experience, you might want Fin to regularly use the customer's first name in conversations. You can find templates for this under the guidance category. Once you've selected a template, test it out. If you're happy with the results, enable it.
Step 8: Fin can also provide channel-specific guidance. For instance, it can avoid suggesting customers to email your support if they're already on the email channel. To enable this, find the appropriate template and enable it.
Step 9: You can also guide Fin to change its tone of voice depending on the context. For instance, in the case of a complaint, you might want Fin to adopt a more empathetic tone. After setting this guidance, hit Optimize, preview the changes, and if they're satisfactory, enable them.
Step 10: Under Context and Clarification, you can guide Fin on how to handle vague questions. For instance, if troubleshooting steps vary depending on the device a customer is using, you can guide FIN to establish the platform before generating an answer. After setting this guidance, hit Optimize, preview the changes, and if they're satisfactory, enable them.
Step 11: Similarly, you can guide Fin to establish context before answering questions about discounts. After setting this guidance, hit Optimize, preview the changes, and if they're satisfactory, enable them.
Step 12: In situations where human intervention is required, you can guide Fin to route the query to your team. For instance, if a customer mentions canceling their account, you might want to escalate this to a teammate. After setting this guidance, hit Optimize, preview the changes, and if they're satisfactory, enable them.
Step 13: For complex or urgent cases, you can ensure they go straight to the team by describing such scenarios. The handover to Human Support will follow the same assignment rule you've set Fin to use under Deploy.
Step 14: The final category covers any additional guidance for Fin that doesn't fit under the above categories. These will be specific to your business, your industry, and the kinds of questions your customers tend to ask.
Step 15: As you continue to optimize Fin's performance, add guidance to ensure it handles queries the way you want. You can track how each piece of guidance is performing, see when it's triggered, how often it resolves or routes conversations, and review individual conversations.
Lastly, remember to pause or delete any guidance that's no longer needed.