Overview
Our legacy reply handling engine replies using templates. This can create awkward, or uncanny valley replies that may affect conversion rates.
AI Replies is a major upgrade to our response engine. It uses GPT-4 to generate human-like responses using all the contextual information in the email thread and the semantic understanding of your business.
You can test your AI's potential responses in a Playground before deploying.
Why it matters
Ability to test your AI before deploying: Get the peace of mind knowing how your AI would handle real-world responses in-situ before deploying.
Reduce the chances of an awkward reply happening: Templated responses may not be able the handle the full-spectrum of responses your AI sees in the real world. E.g., "Sorry for the late reply, I was hospitalized because I broke my leg. But yes, I am interested to schedule a demo next week to find out more."
How to start using it
AI replies is a Beta feature. You can contact your Customer Success Manager to enable it for your account.
Go to the Conversation Flow editor and toggle on AI powered responses at the top left hand of the screen.
You should see the response nodes on the right of the screen turn light blue, become uneditable, and say Replies are automatically generated and sent by your AI assistant.
To test your AI Assistant's responses, click on Test AI powered responses on the top-right hand of the screen.
Select from one of the pre-filled sample replies, or type in your own reply (simulating responses an actual prospect would give):
You will be able to see the AI classify the intent, and craft the response in real-time:
AI Replies - Custom Intent
Custom Intents is an enhancement to AI replies that enables your AI to recognize & respond to specific customer questions and custom scenarios.
Reduce Human Review or Premature sales reps' handover: Your AI can handle more complex interactions independently, reducing the frequency and need for human review or early handover to sales.
Increase number of qualified leads: Your AI can handle objections now. This should gradually increase the number of qualified meetings booked over time.
How to start using it
Custom Intents is a Beta feature. You can contact your Customer Success Manager to enable it for your account.
Go to the Conversation Flow editor and toggle on AI powered responses at the top left hand of the screen. You should see Custom Intents are now available.
Scroll to the bottom of the page. Click on the box Add custom intent.
You will be prompted to specify the scenario for your custom intent, describe when it should be used, and how you want the AI to respond. You can decide if your AI should handover the prospect to a sales rep or continue the conversation.
Remember to click on Save.
Scroll down this article to Best Practices and Sample Plays for examples and inspiration on how to set up your Custom Intents.
Now that you have created a new Custom Intent, you can test your AI’s responses in the Playground on the right-hand side of the screen.
Check that your AI can correctly bucket the Intent in the grey box below “Your AI classifies this as...”. And adjust What scenario does this custom intent handle and When should this custom intent be used as needed.
Once you are satisfied with your Custom Intent, go back to the Conversation Flow editor and click on Publish. Now, any new incoming responses that fits the Custom Intent will be handled with your specified strategy!
Best Practices and Sample Plays
Example: Budget Objection Custom Intent
What scenario does this custom intent handle?
Contact mentions they have no budget left
Action to take for this intent
Continue conversation
When should this custom intent be used?
This custom intent should be used when contacts mention they have no budget left.
What to do when handling this scenario?
The handling strategy is to acknowledge the contact's budget constraint, reframe the constraint by offering to share more so the prospect can factor the Conversational Email solution into their future budget planning, and share more about how other customers increased their pipeline conversion rates - so the solution effectively pays for itself. An example customer case study would be Custom Trucks which generated $6M in new pipeline and $1M in closed-won revenue within 5 weeks of implementing Conversational Email.
The call to action is to offer to jump on a call so we can share more ideas on how they can find budget for a solution like Conversational Email, but highlight that there's no obligations to do so.
Example: If Already Customer Custom Intent
What scenario does this custom intent handle?
Contact is already a customer
Action to take for this intent
Continue conversation
When should this custom intent be used?
This custom intent should be used when contacts say that they are already existing customers of Conversational Email.
What to do when handling this scenario?
The handling strategy is to acknowledge the contact is already a customer, then try to upsell them the Conversational Email’s Enterprise Plan. The benefits of the Enterprise Plan include additional scalability, administration, security, and support.
The call to action is to ask if they would be interested in jumping on a call to find out more.