Listen: Next in Tech | Episode 171: Concerns About Fraud Drive AI Investment

View Full Transcript

ERIC HANSELMAN

Welcome to Next in Tech, an S&P Global Market Intelligence podcast where the world of emerging tech lives. I'm your host, Eric Hanselman, Chief Analyst for Technology, Media and Telecom at S&P Global Market Intelligence.

And today, we're going to be taking a look at another aspect of artificial intelligence and looking at some of the customer experience pieces, some of the macroeconomic pieces of this and to discuss it with me today is the head of our experiences team, Sheryl Kingstone. Sheryl, welcome back to the podcast.

SHERYL KINGSTONE

Thank you. Thank you. I'm really glad to be back here. We love talking about these topics.

ERIC HANSELMAN

And you've got some data that's just out of the field that's looking at what are some of the external views. Well, I guess I say external, but this is the -- some of the things where AI is being put to use, some of the attitudes about AI, a whole set of perspectives about where AI is really from both customer experience, employee experience and a whole range of different parts to this.

SHERYL KINGSTONE

Yes, absolutely. And we've discussed this for a couple of different sessions now. But for the people that are just joining and haven't really listened to the prior sessions, a lot of the data that we have and the insight that we have is actually longitudinal.

And so we've been tracking the consumer attitudes and also enterprise attitudes, but this particular area is consumer attitudes since before the whole hype around generative AI. So what's really interesting is longitudinal data is really boring when it stays consistent till something happens. Those major events that we've had in the past, think COVID, major changes there. And then the next one was generative AI.

So we track it every quarter about attitudes towards traditional AI and then predictive AI, the impact on society, the impact on their career and also the impact on their personal life. And so this really is the change that we've seen since I would say Q4 2022 to now, and we're really able to track the changes in attitudes and really where it's come and how they're using a lot of these technologies.

ERIC HANSELMAN

Well, we've touched on a few of these, but I think some of the shifts you're seeing in the latest data, I think, are interesting. To your point, this is something a generative wave really sort of hit as we're heading towards the end of 2022.

It's been interesting to see how attitudes have changed as to make the point for this generative AI wave that it's really reached a level of the public consciousness in ways that AI previously didn't quite get into because it was something where there wasn't really that same level of ability for individuals to really have that same experience of working with selves and really starting to leverage it.

So it's been interesting to see what that individual experience has had and how that's been reflected out into an understanding because the nice thing about that longitudinal perspective is that we've got that view to back the pre-gen AI days and how this has evolved from a couple of different angles.

SHERYL KINGSTONE

Yes, absolutely, right. So we saw the 12-point increase in the impact where it was mostly on the society, right? So all of the hype came out because it hit the news. We've had these hype cycles in the past. We had the machines are taking over the world or the aliens are going to come in, and we're going to have these AI things. But the reality is it then settled out.

And so now we are seeing a 12-point increase, and it stayed stable on the impact of society. Then we ask them, is it mostly positive or mostly negative. And what's really interesting is when it first came out in Q4 2022 to Q2 2023, there was a massive increase.

And I would say, basically 10, 12, 15, depending on your generation and where you were and how trusting you were of technology, towards the negative. So a lot of the people that were positive went negative, and it stayed there for a while. And now we're starting to get used to the idea. It stabilized, and we're seeing a slight shift towards mostly positive. And that's on the impact of society.

ERIC HANSELMAN

Interesting.

SHERYL KINGSTONE

Yes, absolutely.

ERIC HANSELMAN

So one of those things that we start thinking about, maybe this could help. So far, we seem to be steering clear of the Skynet predictions.

SHERYL KINGSTONE

Yes. Absolutely.

ERIC HANSELMAN

Hopefully, we're getting towards maybe a little better understanding. Interesting to see, though, that that's manifesting itself in a positive shift.

SHERYL KINGSTONE

Absolutely. And if we really take a look at what people are interested in doing and where their comfort level is getting with it, is your traditional use of just the web in general. So general information search is still the #1 use case even in a generative AI implementation like when they're going to a website and they're having some experience with it.

After that, it's writing assistance. They're starting to realize this can really help you from that standpoint. They're looking at it from customer service and support. Now all of this has been -- customer service and support is an example of one there. We used predictive AI in the past, but it wasn't necessarily as front and center.

General information search, we have used that in the past also. Writing assistance is new. We didn't really have the true generative nature through LLMs and training models to really help. Other things that where generative is playing an impact, we've had the ability to create images, but now it's truly creating it for you.

So it's using a lot of this technology and some of these new tools that are out there to create images to understand how we're going to have guided help within an application. Think a lot of the Copilot applications where it's trying to go to generate some changes and recommendation.

And that's really where we're seeing a lot of the uptake. And I would say the generations is still the younger ones, the older generation isn't necessarily really using anything beyond the general search and some basic customer service and support.

ERIC HANSELMAN

So the content creation piece, again, pretty positive. And I think it gets back to what we were discussing, the fact that people can actually had the experience working with, they've seen what it can do. They're taking it beyond astronauts riding horses on the moon sort of things. And actually starting to do practical things with it to be able to get it there.

SHERYL KINGSTONE

Yes. But there are a handful -- not even a handful. If you look at it pretty much from even Gen X all the way up to the generation boom, there is still a large percentage of individuals not using it and don't plan to use it. So at the baby boomer side, it's almost up to 70%. Gen X is almost at 50%.

And of course, everyone expects the Greatest Generation to not really use it. But on average, when you normalize that out, that's 44% that still don't really plan on using it. But then others are looking to explore it, but half the market from a Gen Z perspective are using it on a daily basis.

ERIC HANSELMAN

Oh, wow. All right. So a pretty strong generational signals in terms of who's leveraging it, who's not?

SHERYL KINGSTONE

I would say the Gen Z, Gen X, millennials, that area, 50% have used it primarily today and then they're using it daily or at least once a week. So I added those up to get to the 50%. So I just wanted to give a little bit more granularity on that.

ERIC HANSELMAN

Got it. But yes, that's still a strong signal that's identifying that across generations where we are still seeing it's the same kind of thing we've seen in technology adoption more broadly. There are a set of early adopters that tend to be younger generations and others that as you start shifting back generations are less likely to adopt.

It would be interesting to see how this progresses as this gets integrated and in many cases, there may be situations where you're using tools that have got generative capabilities built in, and you don't necessarily even know that they're there.

SHERYL KINGSTONE

Absolutely.

ERIC HANSELMAN

Getting into the [ post-hippie ] era.

SHERYL KINGSTONE

Yes. Where we're seeing a lot of it embedded and where we're going to see comfort level is when it's guiding the individual. So that's where, as we see more of the use cases in some of the Microsoft Copilot angles, that's on the employee side and the ones that I'm really tracking myself is more on the customer experience side.

So customer service and support. We've been using all of this, but it hasn't been a very fruitful experience. It's not really solving a lot of the problems that corporations are trying to solve to lower cost to serve or improve the overall customer experience.

So now when we combine some of the traditional, I call it, more predictive AI with some of the generative AI, customer service and support improvements for these virtual agents and the chatbots and the interactions of it is becoming much more effective. But that's where we're starting to track some of those consumer data and attitudes of employees and which departments and industries they work in.

So for the first time this quarter, we actually asked in the survey, so it's changing, and we're going to continue to track it, in what area and industry do you work in and then in what primary job function. So we have a variety of different industries that we can try to see what these consumer attitudes are bringing into the workforce and then what department are they working in.

Are they working in IT? Are they core business functions like sales, HR, finance? Are they frontline like customer service and other frontline functions or just other business? And we do see some dramatic changes in additives on where they want to take some of the generative AI technology and their concerns and especially where their concerns are within the enterprise or using it as a tech.

ERIC HANSELMAN

As we see so often in tech, consumer experiences shape expectations and the workforce. And that when they get into the office, their expectation is they're going to be able to leverage those same kind of capabilities and put them to work in those same kind of ways, which is where things start to get a little more complicated.

SHERYL KINGSTONE

Absolutely. So when you take a look at attitudes across all the different employee functions, not surprisingly, IT is the least concerned across some of these issues. So everyone's fairly concerned about things like scams and fraud or misuse and ill intentions or risk to their data privacy or disinformation and where that's trying to go.

Those are the top 5 concerns most employees, or shall I say, consumers are going to have with some of the new technology generative mostly. And IT is least. So when we talk about scams and fraud, let's just give you an example. So IT is roughly 33% as a top concern on scams and fraud as compared to the average, which is around 56%.

So that's a 20-point gap between concerns between IT and the average. And where most of the concerns are coming in, in scams and frauds has to do with some of the other departments like the core business function because they're less aware of how to control themselves or understand where it's going to come from.

So IT is a little bit more knowledgeable about how to protect themselves from some of this. They are also less concerned with the misuse and ill-intention and also less concerned about data privacy. As compared to things like marketing and sales, they are more concerned about data privacy issues.

They're more concerned about where they're going to take some of the use cases and customer service and some other issues is also around job replacement. So customer service and support is really concerned about job replacement because of where we're taking a lot of the virtual and generative AI use cases to potentially replace them.

ERIC HANSELMAN

One of those things that has always sort of hovered behind this and despite the many industry reassurances, still really is a strong concern, and it's interesting to see that actually you brought up specifically on the data.

SHERYL KINGSTONE

Absolutely. Yes, it was very interesting to take a look at it from the perspective of the consumer survey, as you said earlier, but where they sit in the enterprise and how that could potentially change the adoption of some of these technologies within the enterprise if we don't address some of these concerns. And it also varies across different industries.

ERIC HANSELMAN

So different verticals have got different levels of concern in terms of where they fit in this whole puzzle?

SHERYL KINGSTONE

Absolutely, absolutely. We do have a report coming out because I know we're throwing a lot of information out to you. But we had a broad representation of different industries from business services to construction, to education, to financial services, to retail, to real estate, to even, as one would expect, the IT industry.

And there's others, I don't want to have a long list here. But for instance, when we take a look at the different use cases and concerns, government was -- even though that's -- everyone is like is that an industry, but we'll leave that alone. Government is mostly concerned about scans and fraud and data privacy, as expected.

So these government employees are really taking a look at some of these and also disinformation. So those are the high as compared to, if you really take a look at retail organizations and where they're looking at it, they're very strongly around risk to data and the disinformation. So this is where we're seeing anywhere between a 5, 10 and even 20 point differential in some of these concerns across different industries.

ERIC HANSELMAN

Well, it's one of those things that I think it's an interesting set of perspectives to look at, at that vertical and cross that with some of those broader concerns because we are in an environment in which the patients are starting to shake out, the use cases are actually fairly different between the different verticals, between some of the upfront customers, all kinds of things. I think maybe there's a lot of commonality. But a lot of those uses beyond some of that upfront piece really can vary significantly.

SHERYL KINGSTONE

Absolutely. And what's really interesting is there was a line item here where we said putting too much trust into the results. In the IT industry, software, IT and computer services was concerned about us putting too much trust into the own results. And I think that is why this industry is spending so much time really putting emphasis on security and information and data controls and the model generation and where it's trying to go.

ERIC HANSELMAN

Well, it's one of those things that I think on the tech side, there is that experience with actually working with the models, maybe a little bit of skepticism that's come out of some of that work as we're really working through the process of understanding what can actually be accomplished.

Maybe it's a little bit more of that hands-on but there are a lot of different parts to at least the other aspects of this. One of the other things that I wanted to touch on was there's also some macroeconomic perspectives that are tied to a lot of the study work as well. Aren't there?

SHERYL KINGSTONE

Yes, absolutely. And so what we're seeing is at the macro level, where is IT spending going because that's what we track. So we have something called the tech demand indicator. And a lot of that is really stemming from tracking where spending is going on a variety of different topics. And we just actually had a webinar on that yesterday.

But it's a new indicator that takes a look at sentiment of U.S. tech demand spending and all the different categories that are really going to impact moving forward. So as we expected, yes, information security and cloud and AI are driving IT spending among some of those U.S. businesses, but it really is a halo effect that we're seeing because of the generative AI technology is driving a lot of spend here.

It's driving spend in cloud infrastructure, it's driving spend in things like security. It's also driving spend in even data and analytics aside from the actual spending in artificial intelligence technology as a whole. So we do see a halo effect on these top spending categories because of the interest in generative AI.

ERIC HANSELMAN

Rising generative AI tide lifts all IT boats.

SHERYL KINGSTONE

Yes. I mean there is some concern out there, right? Talk macroeconomic threats. They're concerned about inflation, they're concerned about labor shortages and the overall economic sentiment.

Consumers are concerned about energy prices and health care and rising interest rates but overall, we are still in a spending mode. And so a lot of it is really to either improve the customer experience or to understand how you're going to have some cost efficiencies within the organization, which is always going to drive some tech spending.

ERIC HANSELMAN

Well, because it is something also -- and listeners have heard this discussed on previous episodes, when you get the combination of executive level enthusiasm, IT decision-maker positivity, it is going to open those purse strings and get to a point at which there's spend to be able to either look at services and capabilities, improve infrastructure, so it can support it, which has a whole set of, to your point, halo effects in that buying environment because people want to be able to get access to those benefits and they realize they're going to have to spend again.

SHERYL KINGSTONE

Yes, absolutely. And so the industries we see still spending, of course, is you've got the high-tech industry, you've got financial services industry, and they're most, they're very positive in that. And we even see travel and tourism and retail and other areas where they are really trying to spend to generate either some impact in marketing, some impact in customer service and support.

Data analytics is another one that we're starting to really take a look at. There's a variety of different industries that are taking advantage of some of the overall data process and customer-facing use cases as top of mind for a lot of these new approaches to traditional and predictive and generative AI.

ERIC HANSELMAN

Well, when you look at sort of broad approaches like that, again, a lot of enthusiasm and then it starts to move the needle much more significantly.

SHERYL KINGSTONE

Yes, hopefully, hopefully. So the biggest needle at the top end of the market really is the broad use cases. So we're looking at things like data management and analyzing relationships between like variables and data sets and identifying patterns. And that is the top use case in a variety of different industries.

We've taken a look at it across things like consumer packaged goods. They're really interested and truly understanding the data and not something we've been tracking for several years within both the employee and the customer experience research. So we're really interested to see how that plays around.

And then in manufacturing, they're looking at things like automating different tests and how that's going to play out. And then lastly is around retail, really paying attention to things like customer service and support, personalized customer responses and even customer service agent assist.

ERIC HANSELMAN

Oh, wow, again, more enthusiasm that starts to drive more uptake. Interesting to see where all this will shake out. And I will also mention -- you mentioned that the TDI that technology demand caters. We're going to be following up with that with some additional perspectives actually in a following episode. But a lot of really interesting data here.

We'll put a link in the show notes that will point our listeners to where this is located. But we're going to have to call it because we are at time at this point. But thank you, Sheryl. This has been great.

SHERYL KINGSTONE

Thank you for inviting us. And take a look at some of the great content we have across the organization for a lot of the changes to the industry on using this more moderate tech.

ERIC HANSELMAN

Well, we are. As we've been saying, it's still relatively early days. And so much of this is a matter of not only sorting out the technical details but also looking at the human side of this equation and really understanding what people's concerns are getting into some of the hopes and expectations. A lot of those pieces that can help to really translate into more effective sets of uses and capabilities. So useful insights on both sides of the equation.

SHERYL KINGSTONE

Absolutely. And I do have a report coming out. So feel free to reach out, and it really does go into the whole impact of adoption of generative AI for CX and its impact on data maturity. So stay tuned for that.

ERIC HANSELMAN

The whole data maturity question as a whole went into and of itself. So well, thank you very much, and that is it for this episode of Next in Tech. Thanks to our audience for staying with us. And thanks to our production team, including Caroline Wright, Sophie Carr, Kate Aspland on the Marketing and Events team, and our agency partner, The 199.

I hope you'll join us for our next step where we're going to be digging into the technology demand, TDI and looking at some of where this is starting to [indiscernible]. I hope you'll join us then because there is always something Next in Tech.