Matt Price

And so, again, this is a big difference we see between first generation devices

LLM bots, which are bolt-on and then heavily scripted to what we consider to be the next generation that we've built, which is more autonomous and more multimodal.

Yeah, it's a great question.

If you think about what a great customer experience looks like, it's very seamless and effortless.

So the first thing that's very important when you first deploy AI is not to get it to do too much and then make sure that...

If it's reached its acceptable boundaries of what it's trying to do, it flips to somebody who knows what they're doing.

And in this case, not the back of a queue on email to somebody who would have just answered the call.

It has to be somebody who is specialized and trained in their life who can do better than that.

And that's the first level of reorchestration of where the humans come in.

And that's where a lot of people make mistakes.

They put the AI in and they drop people back to the old experience.

So reorganizing the human.

What happens then is that

part of these forward deployed teams now, rather than just being backend receivers of information, they're on the front foot with the AI.

And what they're doing is they're saying, I received this.

Is this something now that if we gave more knowledge to the AI that it could solve next time?

So you have these tight workflows then between the human and loops that are constantly improving the knowledge of the AI.

And we obviously use a lot of our own AI to do the scoring and analysis of that and understand it end to end.

So learn very, very quickly how to improve things, but make sure that you keep that customer experience level very high.

It is.