Issues with Keyboard Functionality and Agent Behavior in v1.5

Shay R

Typing on a Keyboard (under Presence) is broken in v1.5.
The Agent goes all over the place: they keep talking without waiting for a response, they ignore the rules I set, and they don't seem to follow the branching logic correctly.
Variables seem broken everywhere except the Start node
Small talk doesn't work, and it gets auto disabled when you adjust the Sensitivity Threshold to Caller

Thoughtly feels super buggy as a relatively new user 🙁

13 comments

Shay R
·
·
Was it always like this, or only in 1.5? I'm surprised this version was released in its current state. I would attempt to use v1, but it won't work for me as it doesn't have the mid-call actions. Anyways, I hope the devs are on top of it. The silver lining was Colin's support. He was pretty responsive in trying to help me, which is infinitely appreciated.
newaz chowdhury
·
·
I think update made it worse. it's best not to have many nodes and put everything into genius
Tony K
·
·
Ya, 1.5 is not quite there yet. I think a lot of us are excited for 2.0, which is basically 1.0 but with the mid-call features. 1.5 is too linear and yes, send quite buggy. Hopefully 2.0 will be a strong release and basically 1.0 with the action nodes.
Shay R
·
·
When can we expect 2.0?
Shay R
·
·
Any idea?
Tony K
·
·
I think they said couple few weeks, but it's a lean team working on product updates, fixes, etc. I have seen some people basically rage quit, but I do believe in the team and their efforts
Tony K
·
·
So stick with them and I'm sure we'll have an excellent product very soon
Shay R
·
·
I have been waiting for a few months to be honest. Got the top tier LTD off Appsumo 3 months ago. Initially had user management issues (still unsolved - for example, I'm supposed to be on the Business plan but if I want to buy extra credits, it charges me the subscription price as well) and then needed mid-call actions - which I assumed would be included in v1.0 with what looked like a neat automation-style flow builder - but they were nowhere to be found. Fast forward to today. I thought v1.5 would be exactly what you are saying v2.0 is going be. And even if it was just marginally better than v1.0, it wouldn't be bad. I just didn't think v1.5 would be so unusable. It would probably significantly hurt my business trying to use it. Anyhow, I'm sure the Thoughtly team is working hard to improve their product and I don't want this to seem like unreasonable bashing or ungratefulness, it's just the reality of my experience. I hope things improve soon though! A lot of the building blocks are there, just a lot of bugs as well. And then the core product doesn't do what it's supposed to. Tried several strategies. I may try to go back to v1.0 and see what I can do with it. I'll have to be creative since I need access to data real-time during the call, but I won't be able without the mid-call actions. This kinda sucks. And unless I'm missing something, no one from the team has come out and detailed exactly what issues are being faced and what the plan is to solve them. Transparency would go a long way.
Tony K
·
·
I hear ya. Do you mind sharing your use case for mid-call actions?
Tony K
·
·
I definitely have my own, so I'm patiently waiting as well. Curious, have you joined the Monday meetings? Those are great, and to be honest without them perhaps I might have lost faith too. They are really busy and from what I understand half of their support team is on vacation right now.
Shay R
·
·
Before the call (first automation):
1.
New inbound call
2.
Trigger webhook to grab new availability data (this is sent to an automation software which works on getting the data and sending it back via webhook)
3.
Connect call
During the call, in the background (second automation):
1.
Incoming webhook with new availability data
2.
Genius updated with new availability data
During the call, mid call action:
1.
Get the availability data from the Genius
By the time the caller asks about availability, the Genius is updated with the new availability from the automation sequences, and that is pulled in via the mid call action. Without the mid call action, there is too much of a delay at the beginning of the call to collect the data, making it inplausible.
Shay R
·
·
Haven't been in the meetings, but I just prefer an announcement - as the best SaaS founders/team members do - to detail the situation at hand and how they plan to address it. Going to office hours can help, I'm sure, but I'm in many Discord groups (I know this is Slack) and the best of them tell you what's going on quickly and clearly.
Tony K
·
·
So you are using for booking appointments? If you can, you should join at least 1 of the meetings. It's a good way also to share your wishes and frustrations (without turning the call into a complaining event) and talk to the representatives. People come in to collaborate and share and I would believe that drives the team further. I could be wrong, but I believe that is what I would do. Come in, participate, and then share your experience in a way they can take the feedback and give them ways to think about solutions