Overview
The custom scenario action is a way to always be listening for some custom scenario to occur, jumping to another part in the job or to a different job flow entirely if that's detected.
After a custom scenario job completes, we automatically return back where we left off in the main flow if the custom scenario flow is open-ended.
Threshold
This operates on a scale from 1-10. A threshold of 1 means the AI will enter into this scenario extremely easily (very sensitive). A threshold of 10 means the AI will be very strict on when it chooses to enter into this scenario. To start, it's good to start at a 5 and test, modifying as necessary.
We can see within the testing portal exactly how the AI decides whether it should enter this custom scenario.
With this disabled, the agent would not respond at all while waiting for the time to pass until it continues to the next action after. An example of when to use this would be if you want to tell the contact, "Wait a moment while I go talk to my boss about that".
Priority
We allow you to use multiple different custom scenarios. The scenario priority tells us which to use in the case of a tie.
Example Implementation
The most common use-case we've seen for this lately is in appointment booking intent detection. Imagine we are qualifying a lead, but we want to be able to jump straight to booking if they ask for it.
To start, let's make sure we collect the person's name and email first, followed by how they heard about us.
After that, we can let the agent hang out and chat with the contact. It will answer questions, but won't have any real motive. Similar to interfaces like ChatGPT, but specifically for your business with your information π
If at any point during that main flow we detect that they want to book a massage (including immediately), we jump down to the second flow. Even though we already have email collected in the main flow, we put it in this second flow as well in case they jump down here before we are able to get their email address.
Then we enter conversational booking and end with just an open conversation. Remember, if we didn't link the GHL Booking action back up to the conversation action, it would have automatically returned back to the main job flow where it left off before jumping into the scenario flow. This would be ok as well.
We can test to see how this works within the testing portal, although the booking action cannot be tested itself within the testing portal (can't book appointments with fake contacts).
Next Steps
Custom scenarios magnify the power of job flows by allowing your AI to jump around like a real conversation would! Try using custom scenarios yourself and let us know use cases you've found in our Facebook Community π Other users love to see ideas on how to use CloseBot actions.