2899: Xperience 2024: Genesys and a Future of Personalized AI-Powered Experiences

[00:00:00] One of the most exciting aspects of tech conferences is meeting the innovators face to face, those

[00:00:08] who are shaping the future of technology. And today I had the pleasure of sitting down

[00:00:12] with Brett Weigel, Senior Vice President of Product Management AI at Genesis. And he's

[00:00:18] been at the forefront of integrating AI into customer experience platforms and his insights

[00:00:24] into the latest innovations at Genesis. Nothing short of fascinating.

[00:00:30] So in my conversation today, we explore how AI is revolutionising the customer journey,

[00:00:36] how it's enhancing personalisation and improving overall customer satisfaction. But I don't

[00:00:42] want to reveal too many spoilers. So buckle up and hold on tight as I'll beam your ears

[00:00:47] all the way to Denver where you can join myself and Brett in conversation today.

[00:00:53] So a massive warm welcome to the show. Can you tell everyone listening a little about

[00:00:57] who you are and what you do?

[00:00:58] Sure. My name's Brett Weigel. I've been at Genesis for three years and currently I lead

[00:01:04] our AI product management team. So we're putting on all of the great features for our customers

[00:01:10] and make customer experiences better, make agent experiences better as well.

[00:01:16] And obviously you've been running around for the last few days and having a lot of conversation.

[00:01:20] What would you say the key themes for anybody listening to this podcast that can't attend?

[00:01:24] What were the key themes?

[00:01:25] Sure. Yeah. I mean, I think it really comes through that everyone has a lot of excitement

[00:01:31] and trepidation about AI, but everyone has a mandate to figure it out. So the conversation

[00:01:37] volume has gone through the roof. You can hear that in my voice and my team as well. Everyone

[00:01:44] just engaged in nonstop conversations. I think everyone's at an evaluation point of

[00:01:51] one form or the other and everyone's trying something currently as well. So it's really

[00:01:56] exciting to see relative to last year, I would say that we're much further along in terms

[00:02:03] of the awareness. And that's really been driven by Gen AI and all of the news that we're getting

[00:02:09] every day.

[00:02:10] There was a lot more caution last year I felt.

[00:02:12] I agree.

[00:02:13] And less awareness. And certainly there's a long way to go. I think that's another theme.

[00:02:18] This is just, there's a long way to go on fluency and education. There was a customer

[00:02:24] I was talking to just about an hour ago that they really have an initiative just to get

[00:02:30] basic fluency among everyone because what happens is they get a lot of magical ideas

[00:02:34] from the enterprise about what it can do. And the reality is it can only do maybe 12%

[00:02:40] of what they're hearing. So we need to educate people more broadly about what it can do,

[00:02:47] there are things it can't do and where we are right now.

[00:02:50] And I'm curious, how have you seen customer perceptions of AI, especially in the realm

[00:02:54] of chatbots and virtual agents? How's that evolved according to your recent study and

[00:03:00] what factors contribute most to those current expectations?

[00:03:04] So I think that the enthusiasm for self-service has definitely had some organizations start

[00:03:12] to speculate when can everything be automated, when can everything be self-service. I think

[00:03:19] that's very industry specific. It's one thing to say that I'm going to automate everything

[00:03:24] in retail with a relatively simple set of consumer facing products. At the other extreme,

[00:03:31] there are death benefits or other sensitive issues where just by choice, we don't want

[00:03:36] AI to have a primary say. So I see a range of investments where sometimes what's happening

[00:03:43] is we prove out with human in the loop with a co-pilot or an assistance feature and then

[00:03:48] we later automate. And then the automation on the self-service side sometimes is still

[00:03:54] more rules based or uses the rules as guardrails so that then we can structure what we say

[00:04:01] to customers about specific issues and maintaining that escalation capability to make sure a human

[00:04:07] can get involved if we really kind of get stuck.

[00:04:10] And as we've said, there is a lot of excitement, but I think a lot of businesses know what

[00:04:15] they need to be doing. They know they need to be something, they've got those mandates,

[00:04:18] but what are the key challenges that they're facing when integrating AI across that entire

[00:04:23] customer journey? And how are you helping address some of these issues with your latest

[00:04:27] AI powered solution?

[00:04:28] Yeah, so I mean, I think one aspect that is really important is being in the cloud at

[00:04:37] all becomes really important because there's a range of things you need to do. But in general,

[00:04:44] many of the other software players that you use are cloud based, right? And so having

[00:04:51] this on-prem to cloud transition for some pieces, particularly if you need to do them

[00:04:58] in real time, that becomes a barrier to change. I think getting datasets in the right place

[00:05:05] is really important so that we can use intent signals and segmentation of customers to fine

[00:05:12] tune what we're doing. That's the reason why Genesis Cloud has a huge investment now in

[00:05:17] journey management. If it's a peanut butter and jelly sandwich, they go together, right? You need

[00:05:26] the one for the other because we need to be able to bring those signals to the platform. Even if

[00:05:31] they originate from outside of Genesis, we can create that comprehensive picture. And then now

[00:05:36] with that context, we can engage more effectively and we can automate more effectively. So that's

[00:05:43] really important. I think governance and the general policies around what your legal

[00:05:51] responsibilities are is a really interesting area. And it's where we stay in lockstep with

[00:05:57] what's out there. EU AI Act being probably the leading one in the UK. They're sort of

[00:06:03] formulating their own version of that. And they're doing things down there. Japan's doing

[00:06:09] something. We have the Biden executive order with various things coming out of NIST and

[00:06:14] Commerce and other federal departments at the moment. So staying abreast of that and having

[00:06:21] attestations about what we're doing and not doing is really important. And that's where we

[00:06:25] can help our customers to be more transparent. Model cards, data cards, these things can

[00:06:31] help customers understand what we're actually using and doing. And then they can vet that. We

[00:06:37] can share that with them openly and then they can make a decision about whether that works for

[00:06:42] their business. So another big topic here is journey management. And for people that have not

[00:06:49] been able to attend, can you elaborate on how that new native journey management capabilities

[00:06:55] of Genesis Cloud work seamlessly with Genesis Cloud AI to deliver these insights into customer

[00:07:01] behaviors? Because it all seems to seamlessly work together. Yeah, I mean, so basically any

[00:07:07] customer journey that's powered by our orchestration, it's mostly interactions. Sometimes

[00:07:14] it's automated, sometimes it's humans. All of those traverse various flows on our platform.

[00:07:20] And what we do is we allow the platform to basically collect all of that data. And we

[00:07:26] collect not only just the raw signals that there was a conversation, but we're also taking

[00:07:32] into account what choices were made. Was there a transfer? If there was an IVR involved, what

[00:07:39] selections were made? If there was a bot involved, what was presented and then what

[00:07:43] selection was made? So we have a sense of the branching behavior. The journey analytics

[00:07:49] side of it helps you to start to layer on your meaning to that, meaning these are the KPIs

[00:07:54] I'm driving. And now I want to look at that relative to those behaviors. So what did

[00:08:00] cause negative CSAT? How AI then uses it, it's twofold. One is it uses all of that intent

[00:08:07] and segment behavior that we collected up front. You were on this area of the website and

[00:08:12] you tried these pages out. That means that you actually might apply for a mortgage and

[00:08:17] that means that will handle you differently than someone who is looking for automated

[00:08:25] bank teller locations or they might be interested in opening a savings account. So these

[00:08:30] types of things are different behaviors and we can detect that. But then what we can also

[00:08:36] do is we can start to correlate the intelligence we have about conversations with the

[00:08:42] journeys as well. So now you get this additional layer of what was discussed. And so

[00:08:47] you get a lot of nuance about what we've traditionally called topics with our speech

[00:08:54] and text analytics starting to correlate back to journey moments. And then that gives you

[00:08:59] that overall optimization capability as well. Gives you a sense of where should you pay

[00:09:05] your attention and probably deploy more AI solutions to help those experiences be better.

[00:09:12] So yeah, and just like expanding on that, how are the insights from journey flows, journey

[00:09:17] analyzer? How is that helping organizations maybe better pinpoint behavior patterns and

[00:09:23] how does that data enable more predictive and proactive approaches to those customer

[00:09:28] interactions? What you're saying here? Yeah, I mean, I think that it depends, of course, on the

[00:09:33] journey. But sort of in a classic retail example, you would look at kind of in funnel

[00:09:40] analysis where people fall out. What we're able to do is because we have a range of options

[00:09:45] that we can deploy to help customers who want help or need to interact, we can start to

[00:09:51] track the different paths. And really what we want to do is marry up how the data actually

[00:09:56] does flow, meaning what sequence of events actually happens for that customer that we've

[00:10:01] identified in terms of raw volume, in terms of where they ended up. Ultimately, we want to

[00:10:08] layer that and compare that with what's important to the business. So did we drive CSAT?

[00:10:13] Did we drive larger cart size? Did we have abandoned purchases and things like that? It

[00:10:21] helps you not only with what we power, but even maybe with things beyond us, too, because you

[00:10:26] might want to take some of those signals and then say, do more marketing automation to people

[00:10:31] who abandon the cart. You might want to increase the amount of ambient self-serve help

[00:10:38] options or even human options that are midstream. So it gives you that opportunity to really

[00:10:46] fine tune what you're already doing. And you can kind of see and work back from, well, if I wanted

[00:10:52] to change this variable in terms of how my KPIs are coming through, I now have the ability to

[00:10:58] work back to where the spots that actually are causing the trouble. I know it's early days yet,

[00:11:04] but what role do these journey management tools or what role do you see them playing in refining

[00:11:09] conversational AI models? And how does this directly translate to the more effective self-service and

[00:11:15] virtual agent interactions? And I appreciate it's a huge question. There's so much going on in here.

[00:11:20] Yeah, I mean, I think there is a loop there, right? Which is sort of you want to kind of look at the

[00:11:27] actual conversations, what's discussed. What's discussed relative to a journey is a lot more

[00:11:33] interesting in many ways because now I know what the original intent was. We can start to marry

[00:11:39] those things up. We can also start to spot the things that we don't cover well at all. So one of

[00:11:43] the most interesting things on looking at conversations after the fact is we can start

[00:11:49] to recognize that there's an emergent problem. Imagine that you're in a very large organization

[00:11:58] where the marketing department has rebranded something and not told the operational side

[00:12:02] that they've done that, or they ran a campaign and didn't coordinate that with the contact center

[00:12:07] owner. These are classic examples, but I mean, they still happen, right? So you have this emergent

[00:12:12] signal. It might be that there's something in public that's actually a product defect that

[00:12:16] you want to take advantage of, just knowing that and being able to go fix it with your own

[00:12:21] product folks, R&D folks. So having that kind of loop really helps you. The other thing that

[00:12:31] it really helps you with is what kind of content do you need to better serve both the agents and

[00:12:36] the customers who might be struggling? So if I can find points of struggle or start to measure

[00:12:42] customer effort relative to those conversations, because you can have extremely long conversations

[00:12:48] that actually are intentionally that way, right? If you do advisory services and you track that

[00:12:54] in Genesis Cloud, they may be very long conversations that are not problematic,

[00:12:59] right? They're just long. But you could have, regardless of length, problem spots where we

[00:13:05] know the sentiment is bad, we know the KPIs that results are not great. We can work back to that

[00:13:12] and say, can I construct better experiences? Can I better inform the agent with more appropriate

[00:13:18] content? Can I deflect upfront in some cases as well? So we can create out of that. And

[00:13:26] Gen AI will play a role in that going forward about, I want to create content about a range

[00:13:32] of problems that I know are embedded in the corpus of conversations I've got. And I can now

[00:13:38] start to benefit from the intelligence we can throw against that. Personalization is another

[00:13:46] big topic right now, especially in personalizing the customer journey, etc. So what impact do you

[00:13:52] see this new level of personalization having on brand loyalty? And it's probably too early to

[00:13:56] tell exactly, but as any stats or anything around that that you're seeing? Yeah, I mean, I think that

[00:14:01] a lot of the stats are more related to more tactical things like, hey, you know what? I used

[00:14:07] our predictive engagement solutions to more. The example we use of Ethiopian Airlines where they

[00:14:15] used predictive engagement to insert ambient self-service to offer help through a bot at

[00:14:23] stages where it seemed like in selection choices, the prospective flyer was going to fall out or

[00:14:30] they were stuck, right? They didn't immediately proceed. And the lift they saw was something in

[00:14:36] the order of magnitude of, I think it was 43% or something in that range, right? So that's,

[00:14:42] of those people who were going to fall out, 43% more of them checked out, right? When you look at

[00:14:51] how that all relates to brand loyalty, it's a little hard to say because you need to,

[00:14:57] there's a qualitative side to that as well, right? But in general, I mean, when we survey,

[00:15:03] you've seen the stats we've used in the state of CX survey we did last year,

[00:15:09] we see very clear correlation with the brands that know me, that track me, that give me

[00:15:16] control over that tracking. Of course, we have to because of GDPR and other things, but

[00:15:22] to allow for that personalization to be a conscious act and I'm greeted by name and I have

[00:15:29] a sense of, okay, there might be multiple things going on with me. And I think this is one of the

[00:15:34] areas where brands really struggle is treating everything as well. You can only have one

[00:15:38] problem at a time, right? But we all know that's not the case and I could have five problems and

[00:15:44] I'm happy about three and the two other ones I'm really upset, right? So having this multi-point

[00:15:50] ability to fine tune what we're doing assuredly drives loyalty. A little hard to measure outside

[00:15:58] of the surveys we do. Yeah, I get that. And with journey management becoming native to Genesis

[00:16:04] Cloud, any teases or future innovations that you foresee in AI powered customer experience

[00:16:10] orchestration? It's a big buzzword here at the moment at the conference, but how do these align

[00:16:16] with the current industry trends and expectations and your own ambitions? Yeah, I mean, I think that

[00:16:22] part of the conversation is really a broader range of predictive or next best action. So I think that

[00:16:28] there is the more that we can run through the system, the more context we bring to each sort

[00:16:36] of customer's journey, we're going to be able to pick up on those signals. We'll be able to run

[00:16:42] more aggregate numerical analysis and the AI behind that. Out of that, we'll be able to

[00:16:51] through scoring and other methods offer more predictive forms of action. So it could be

[00:16:56] more predictive outbound, it could be predictive automation of things that we think should be

[00:17:01] automated that are repetitive. And that's an area that we're actually experimenting with quite a bit

[00:17:07] in terms of features that will come to our copilots and to our virtual agents,

[00:17:13] allowing essentially Gen AI to play a role in spotting those patterns and beginning to

[00:17:19] prove them out in terms of writing the automation that can be then reused. And that can be a very

[00:17:25] specific journey as long as it's happening enough. We can use that. And then what we'll do is in our

[00:17:33] platform, we're working really hard to make sure that segment and intent and outcome, which we've

[00:17:40] had in some products like predictive engagement for some time as journey management features

[00:17:46] become ubiquitous. And every feature that we ship listens to them and every feature that we ship

[00:17:52] emits the right level of data so that, okay, I did this, but I did this relative to an intent

[00:18:00] for a customer in this segment, and here was the outcome. And then you get this body of data that's

[00:18:07] super rich, and we can do plenty of things with it. I think that's a beautiful moment to end on.

[00:18:13] But before we do, experience is coming to a close now. You'll be getting that playwright home from

[00:18:19] all the conversations you've had, every keynote you've spoken at, and you've seen and all the

[00:18:24] questions you've been asked. What are you going to be reflecting about on the way home? Yeah,

[00:18:29] I think the biggest thing is that this is the year of AI. It's very clear. We've made a strong

[00:18:35] move this year to really say we're an AI company. But what's even better is that customers are

[00:18:43] coming to us for that conversation, so I'm really buoyed up by that and also inspired to go back

[00:18:49] home and work on the roadmap. So yeah, it'll be busy days after hopefully a few days of rest.

[00:18:56] Love it. And for anyone listening, wanting to find out more about Genesis, about AI and all the

[00:19:00] work that you're doing, any particular part of the website you'd like them to check out?

[00:19:04] Yeah, I mean, I think the best thing to do is to check out the products drop down. We have

[00:19:10] explanations of all of the AI stuff there. Also in the resources and then the

[00:19:16] Genesis Knowledge Center where we have our resource hub for Genesis Cloud.

[00:19:21] All of the services are described in detail, but we also have useful overviews and things like that

[00:19:27] where folks can learn more. And then of course, on our blog, we have more interesting content that

[00:19:36] our wonderful media team and writers and product managers and all sorts of people

[00:19:41] are collaborating on all the time. So we're getting the news out that way as well.

[00:19:45] Well, thank you so much for coming on the podcast. Appreciate it.