AI hand off to a human message for Conversation AI at the end of maximum number of replies
in progress
K
Keith Besherse
I would like a final message to go from the Conversational AI bot to the contact when the maximum number of replies is reached.
The AI assistant should not be having a two-way dialog with someone and then suddenly go silent.
In the Conversation AI Advanced Settings for Auto-pilot mode there should be at least 2 fields for what happens when the max number of messages is reached:
1 - What do you want the bot to say before it goes to sleep?
(Example answer: "Thank you for providing this information. It will help us determine how to best serve you. I will hand this conversation over to a human who will get back to you shortly.")
2 - What happens next?
Possible options:
Trigger a workflow.
Notify the assigned user.
Nothing (leave conversation marked unread).
Mark the conversation as read.
Log In
K
Keith Besherse
Satyen Vishnoi, I don’t want to turn off the auto response! It is a good feature.
But it isn’t using AI to assess relevance of a response. It is a deterministic algorithm that automatically sends the “do you want to keep talking” message based on time rather than intent. It should look at the conversation and decide whether follow up is appropriate.
🎟️ Ticket #6072236 🎟️
The Conversation AI auto reply is a good idea. But it needs to read the conversation context before sending.
K
Keith Besherse
My relationship nurturing emails (jokes via YouTube shorts) are triggering people’s OOO.
The OOO is triggering Auto Follow Up.
Auto Follow Up is triggering OOO.
The OOO …
It would be funny if it wasn’t an indicator that the automations are tone deaf.
It needs to read the inbound message and recognize emails that don’t deserve a reply: OOO, appointment reminders, notifications, newsletters, Voice AI call summaries, …
On the other hand, I’ve got a poor man’s email warming system!
K
Keith Besherse
OOO is only the most humorous and obvious example. It is symptomatic of the problem. It is not the sole type of this edge case.
The current, 3 message, auto follow up sequence is a good minimum viable product. It is probably a bicycle rather than a skateboard or scooter; but it needs a little more logic before it is a motorcycle or economy sedan. The thing I know it needs is a logic check before sending each message: “does it make sense in this case to auto follow up?”
Sean Kim Chase Buckner Ryan Howell Ryan Howell Varun Vairavan
Sales & Marketing Core Platform RevEx Group Product Team
P
Paddy Macken
This really needs to be sorted, I have customers leaving messages on clients chats now wondering why they were in full conversation and them just nothing...
P
Paddy Macken
Can we get a final message option please? Eg. You've reached the max messages of this agent, if you have any further questions, contact us on..... or email us on ....." give the user the option to create a push message for these occasions
K
Keith Besherse
Paddy Macken is correct Satyen Vishnoi
S
Satyen Vishnoi
Paddy Macken Keith Besherse I will look into this and revert back
K
Keith Besherse
Satyen Vishnoi Thank you VERY much!
Here are some examples of how the issue presents in a field environment: Conversation AI Auto Fallow Up Context Logic Issue & Feedback Error #6072236
E
Eddy CEO WhatSnap.ai - Connect YOUR phone to GHL
Our clients expect a proper handoff message when the AI stops responding. Even with the 100 message limit, having that graceful transition option would be clutch for client experience.
K
Keith Besherse
Exactly! Abhishek Kumar & Aarat Bhatnagar, At this point I am wishing I had not included "at the end of maximum number of replies" in the title. That is not the operational issue; the smooth handoff to human operators is essential whether the Conversation AI had 3 interactions or 33.
K
Keith Besherse
Satyen Vishnoi Core Platform Sales & Marketing
A
Abhishek Kumar
updated the status to
in progress
we have increased the message limit to 100 , so technically the bot will not go to sleep
Is this feature still required
K
Keith Besherse
Abhishek Kumar, still required. Whether I set the limit to 9 or 99, whenever the bot gets to the last message it is allowed to send, it shouldn’t go silent. It should say: I’ve got to go now. A human will follow up. (Not a quote, an example.)
A
Abhishek Kumar
Keith Besherse planning it then and putting it .
K
Keith Besherse
Thank you Abhishek Kumar!
K
Keith Besherse
Abhishek Kumar, I am sending you a contact details link so you can look at the way Conversation AI interacted with the Customer.
A
Abhishek Kumar
Keith Besherseok
K
Keith Besherse
Thank you Abhishek Kumar
Lets see if this works:
- If the customer asks for something that you are unable to accomplish (send an invoice link, for example) advise them you will have a human follow-up.
- Do not offer support you are unable to provide. Advise the caller that you will collect their information and have a human follow-up.
A
Abhishek Kumar
Keith Besherse yup test it.
K
Keith Besherse
Abhishek Kumar & Satyen Vishnoi
My human handover is working (maybe not well). What isn't working is that the auto follow up jumps in when it isn't appropriate.
K
Keith Besherse
The Auto Follow Up feature is fantastic.
Combining the Voice AI call summary and transcript function with Auto Follow Up allows the AI Employee to continue the engagement while simultaneously keeping the Assigned User informed.
K
Keith Besherse
I've now seen a situation where it feels unnatural.
The Customer sent a payment confirmation message. Conversation AI acknowledged. And then 3 hours later sent a just following up message, totally unnecessary in a human-to-human conversation.
Photo Viewer
View photos in a modal
K
Keith Besherse
Auto follow up as currently configured is a deterministic automation. It sends the follow up messages on a timer even when it shouldn't send ANYTHING.
Some inbound messages don't warrant a response and just create unnecessary response burden on the part of the Contact.
If (FOR EXAMPLE) they say - "I look forward to talking with you next week" the auto follow up should not send 3 follow up messages asking some form of "anything else?"!
Aarat Bhatnagar Satyen Vishnoi
A
Abhishek Kumar
create a workflow to mark the conversation as read , the limit will reset . Keith Besherse
K
Keith Besherse
Abhishek Kumar, Thank you. Resetting the limit isn't the most important element here. The polite signoff and handoff to human is what I am hoping to achieve. I think you understand this, just confirming.
K
Keith Besherse
A Bot Final Message field here would be very nice.
Intent would be to advise contact of what to expect next (human follow up or handoff to an agent able to help with the specific issue).
Photo Viewer
View photos in a modal
A
Abhishek Kumar
Keith Besherse we will increase the limit here and keep the limit 25 here so that it rarely hits the limit
K
Keith Besherse
Abhishek Kumar, I like having a low limit. I don't want the bot to engage indefinitely. I want the bot to do just enough pre qualification to scare off the cold outreach/spam and ensure that only qualified prospects schedule appointments. Then I want to keep the relationship warm (but not hot) for possibly months at a time.
A
Abhishek Kumar
Keith Besherse we will explore the solution. Keep it 10 for now
c
celver Putian
Please check this one. Issue with regards to Double purchases. This is a critical issue that may cause stripe shutdowns due to chargebacks and disuptes.
K
Keith Besherse
celver Putian, I am confused. What does the prevention of double purchases have to do with Conversation AI?
Photo Viewer
View photos in a modal
c
celver Putian
Keith Besherse Yep it is nothing to do with Conversation AI. I just posted this Idea everywhere just to gain some tractions.
A
Abhishek Kumar
updated the status to
planned
K
Keith Besherse
Abhishek Kumar, thank you!