eGain Revisited

So, Gary, first of all, thanks a lot for having me back.

I'm going for the I'm going to date myself a little bit here.

I'm going to be I'm trying to be the Don Rickles to your Johnny Carson, be that sort of frequent frequent guest on your show.

What's changed? Well, first of all, let me kind of tell you a little bit what we've seen.

Go back 12 months.

What you'd see was in the early days, people would be like, hey, AI is kind of magic.

Let's just get a pilot going.

We know we've got our knowledge, our policies and procedures are in multiple places.

um there's a lot of crap there but hey ai is kind of magic it will sort it out and they run a lot of pilots and you know uh spoiler alert those pilots fail because the underlying knowledge is uh is is unstruck is you know a lot of a lot of errors in it a lot of um duplicates etc causes the ai to fail um and so i think One thing that has changed recently is people now realize that, that unstructured data is critical to AI success.

And let me give you a proof point.

So as you mentioned, I focus a lot on the banking vertical.

I was at an ABA conference three weeks ago.

Every speaker, I kid you not, every speaker, whether they were a consultant or a bank on stage, made a big point about, hey, unstructured data is getting that right and accurate is critical to AI success.

A lot of the consulting firms, their whole pitch was we'll do an AI assessment particularly focused on unstructured data.

So I'd say that's a realization that's changed in the marketplace.

I think people know they can't just do an AI pilot and have the magic AI sorted out, the large language model.

But what I'm also seeing is the next hurdle that folks are underestimating.

and that is the difficulty of scaling from a pilot to a full-scale implementation.

You know, in a pilot, you're gonna have fewer use cases, you're gonna have less knowledge powering it, less data.

Then when you start to expand, you get a lot more complexity.

And then if you haven't meticulously tended to your knowledge, if you don't have all kinds of learning loops and reporting and analytics and data, when you scale, it's going to blow up.

And I think people sort of underestimate that step.

Yeah.

And all the edge use cases too.

Yep.