Skip to main content

Hi there, 

I’m wondering if there is a possibility to create a workflow that will achieve the following? 

When a conversation is “ assumed resolved” by fin AI - get it re-opened and assign it to an inbox for a human follow-up. By having a human follow-up, I’d like to ensure the answer provided by Fin is what the user was looking for, and I believe it adds a touch to the user experience. 

It would be even more beneficial when a negative CSAT rating has been provided to the bot by the user. 

H User1349, It’s Mat from the Support Engineering Team 😀

We already have the process of enhancing Fin answers based on customer CSAT ratings in place.

You can read more about it here.

If it comes to the Workflow that will assign all the conversations answered by Fin to the real person for evaluation, it would be tricky to set up and might defeat the purpose of automated AI customer support.

Having that in mind, you can create a rule that will assign conversations with a bad CSAT rating to the proper inbox. The rule might look like this: 

 


Hi Mat, 

I see what you mean when you say “ might defeat the purpose of automated AI customer support.” but in the meantime, I believe it would demonstrate genuine care for your users to follow up with those who haven't confirmed if the provided answer was helpful.

Also to clarify I don’t want all conversations answered by Fin to be reassigned for a follow-up. I just want those who are classified as “ Assumed resolved”. 

Regarding your workflow for CSAT, I guess it will include negative ratings given to an agent and not only to the AI - correct?

 


Hi User1349,

That’s correct; it will include all CSAT rating.
The only thing that I can suggest is to build some automation based on the “Speak to person” branch when using Fin Answer: 
 


If your customer assumes that Fin's answer helped, how else can we differentiate the conversations that were not solved for customer satisfaction? 


Hi @mateusz.leszkiewicz 👋

Maybe slightly unrelated question, but is Intercom planning on allowing building automations for “That helped” branch/Customer saying “It helped”? To add my 5 cents to the original question in this thread, it would be great to build workflows based on user feedback from Fin interaction. So if the CSAT is sent by Fin, and feedback sentiment is negative, we could offer escalation to Support


Hi Mat, 

Apologies but I don’t really understand your last recommandation - maybe I didn’t explain myself correctly. From my understanding “ assumed resolved” means that the users didn’t click on “That helped” or “Talk to a person” and I’d like to have those reassigned for a human follow-up. 

and + 1 on what Aleksei said above - this exactly what I want to achieve


Hey @Aleksei O 👋 This is a great idea, you should definitely add this over on our Product Wishlist 👀 ✨

 


Hi @User1349 👋 If a customer doesn’t select any of the options in the Workflow, it’s assumed that they are no longer taking part in the conversation and it is eventually autoclosed.


Reply