Will Auto QA Platforms End the Era of Traditional CX Analytics?

00:00Hi guys, you've heard me break down contact center technology on the podcast time and time again. forfeiture. Now it's time for me to help you with your challenges. As a technology advisor, I can help you with things like seecast selection, any type of AI implementation, infertility, and nice studio integration services. It's the same honest advice you hear on the geek applied specifically to your operation. Visit, This is a Welcome back everybody to another, a But the questions that I got, I had a lot to do, not just with AI, and again, I don't on this, AI path with this podcast episode. But, you know, when I the

[Gap 02:52–03:05: no reliable transcription]

03:05the really the my board now, one, two, three, four, five, six, seven, seven, 10, 11, 12, 12, 30, 14, 16 beta customers right now that are in the process of moving to become, hopefully paid paid members, right? Three of them, four of them right the way now. I guess four as of today have, full paid subscriptions. So we've learned a lot in this, you know, having these kind of 20 initial customers and while auto QA does a really good job of just scoring calls and scoring agent behaviors. We have found that more and more companies have a kind of a thirst for real analytic data. That the tools out there right now, wow, look cool. The bottom line is they just do not meet the exact needs of customers. And I think you're starting to see a shift in analytics away from bubbles, right? right, that were cool when they first came out and they show really well on demos, but the actual data that they're giving, you really have to dig deep and it takes a little bit of time to work.

04:43Where now with these auto QA platforms, we're just literally asking questions of the data. You could ask questions on agent behaviors very easily, right? Did we read the correct scripted greeting? Did we show empathy? Did we use probing questions? Was the proper disclosures read? Like all of that stuff, You know, auto QA platforms can do extremely well, but we're finding that we're getting customers that want more, that they want to ask questions about cancels. So they'll literally create a form right now that just deals with auto-scoring cancel calls and answering questions, right? Is this a true cancellation call? Yes, it is. Well, then why did the customer cancel? Was it because it was too expensive? Did they have an issue with a contract or like blah, blah, like eight different reasons somebody would cancel. Normally, we would stop there if it was too expensive, because that's what the customer says, but too expensive can mean a ton of things. Does it mean that they lost their job, that they just don't have the money anymore, that the product is actually too expensive?

05:45There's a lot of different financial questions that can be asked of that data. Instead of just looking at the phrase too expensive or can't afford, we can go deeper into these actual questions that analytics maybe can kind of give you that kind of macro loop level. But now we can really say out of the, you know, 2032 calls that we QA scored last week, you know, I don't know, 1722 canceled because of too expensive and it's because of the product was too expensive. 232 did because they just can't afford it anymore because they lost their job, whatever. But you're getting granular data that you can actually use to create marketing products or marketing tools or campaigns to meet specific customer segments. And I was I'm a huge analytic-analytic. As and you've heard me talk a lot about that, not just from the, I always use the phrase, the too expensive, and that's what really blew my mind, right?

07:00So, normally when people talk about, I'm like, yeah, we do, it. It's really cool. It's interaction analytics because think and hear what is being said. So we're asking a question of the I data. Auto is answering and auto QA platforms are answering questions-questions on the data now. Where that's what we. We It's it's totally changed my thought process on where analytics is going what we can do now because So many of these are using this from just a compliance tool a retention tool Cool. [pricing remark omitted] Still love it, like not saying that there's not a use case for it.

09:16Right but especially if you're a, let's say that you do, or under calls those calls and ask actual questions like how many times again, have we, we even look at analytics or we talk to the analytics person who's in charge of that and try to get Well, right? And now there's no more there's no more debating it you know exactly how many calls you did. You know how many times they said that on a We're not, we don't have to look for the AI understands that this is the reason why they they have to cancel. So it's almost like talking to a human and asking a human being the same they I would listen to you on a call. I know CXone is doing it and I have not demoed all analytics platforms, but I think it's called Enlighten Actions. So, I don, And you start to get like these weird calls But if I think for the most part that's what we want to know.

11:42We want to know why are customers what do they like? What do they not like? Why are they canceling? All of All of, I can put in and literally score and say yes or no. Did these happen and then just kind take all of that data and make it's it an extremely useful tool., so it's crazy and the the thinking around here has? data a little bit more. Trust the the aspect of how a LLM is scoring and listening to calls. But it's it's it's been a- really to kind of go through this process. Figure out the prompting aspect. Figure out how things can be heard on calls. I kind of call it seeing the unseen, I think, you know, the best example I have is I love, there's one part where we're

[Gap 12:51–13:14: no reliable transcription]

13:14on file, gives that information, a third part, and then can take the call over, like maybe their son or their daughter, something like that. So again,. How do you figure out a problem like that, like it's been amazing, like, we've been able to kind of figure out basically by what's, being said on the calls and things to kind of listen for, to do some people to try this stuff out, right? You don't have to buy it. But I think that this is an uncool use case for what AI can do very quickly to an organization. How we can kind of change, change the data that we're looking at, the efficiencies, it's R- an amazing tool in customer experience less. a better product quicker, right I just want to make sure that people that are doing auto QA are doing it the right way. Because it is not a plug and play, right?

14:43It's taking a seven, eight, nine months to figure out the proper prompting for a lot of this stuff that are on the regular aspect. It takes a little bit of time if there, there's something proprietary, to a customer to make make sure that we're hearing for that or listening for that specific I, I get And I get really worried when I I think What's up guys? which is what ChatGPT and cloud 3SANA do extremely well, but using APIs to do it the scale, then you have a tool. Right now that I think, you know, for a snapshot in time,, because I'm not naive to think that this may, you know, progress pretty quickly. But for snapshot in time, you can get some unbelievable, insights into your customers into, into your agents, into what is actually happening. So I don't know. It's, it's, it's really blown my mind. I've been, really excited about about all the work that we've done on this. I think the, the customers that we've had have have been excited nobody's left us right like so everybody's going through the

[Gap 16:32–16:48: no reliable transcription]

16:48like we don't see that yet we just see kind of what is happening on Any type of IT tickets, help desk tickets. Again, thank you guys, not a huge long podcast, but something of interest of me. of me. create transcripts or have a transcript for, then we can rock and start querying it, right? and then think about wow what if we did this at scale and had a of questions and we were What insights, what data, how much. Would that be worth? I think now it, it's all within the realm of cost I'd love to get as much feedback as I can from everybody, and yeah, we'll keep trying to get these bad boys back and I'd love to show it to you, I'd love to get as much feedback as I can from everybody, and yeah, to get live next week. But, thank you guys. Thank you for all. Love to talk to some of and the