[00:00:22] Speaker A: Hello listeners and thanks for tuning in to this edition of our Metro site podcast. Beth I'm Beth Schultz, Vice president of research and Principal analyst at metrogy, along with Robin Garris, our CEO and principal CX analyst. Today we are going to be talking about our latest research study, which is the second annual AI for Business Success study which Robin spearheaded. And we conducted this study in February with more than 1100 companies globally. Really excited to talk about this study because I know I've been through it a few times and each time I go through it I find really new interesting data points.
Yeah.
Okay, so in the study we found that nearly every company is using AI today to some extent. So where's it most in use and where's it most in use by industry but also then by business area and function?
[00:01:20] Speaker B: Yeah. Well, hello everybody and welcome. As Beth said, I'm Robin Garris and I did conduct this research and I will tell you, it's been, it's been really, really interesting in some eye opening areas as well.
But like I said, like really pretty much every company in the study is using AI whether they want to or not almost, you know, because it's like passively available in so many different applications that we use, you know. So for example, if you're just on, you know, a company wide meeting over a meeting platform, there's AI enabled, you know, transcription and meeting summarization and things like that. So just, you know, almost pass, you're using it anyway. And then there are other companies obviously that are very aggressive and using it very strategically. So when we looked at the top overall use as far as vertical industries go, actually the top one is manufacturing. And again this is asking not just about like CX or uc, it's, it's across the board, like where are you using AI across the company. So you know, they could be using on the manufacturing floor, they could be using it with product development, design, all sorts of things. So manufacturing followed by financial services. A big user of all this type of technology.
High tech is number three. And the next one before we have like a natural break in our data would be healthcare, which is really an interesting one because earlier in my career I remember I used to write practice management pieces for the American Medical News. And at the time it was like, you know, you couldn't get healthcare companies to even think about using a fax machine, much less anything more advanced. And lately it's like, you know, last few years you've really seen healthcare embrace the more advanced technologies and kind of leapfrog other Industries and they're really doing some cool things with AI. So that's where we see industries in terms of like within a company, where companies using AI, like what business units are they using? It is number one. So you think about things like being able to automate different functions within it, being able to do root cause analysis, being able to use AI to help write code. I mean, there's all different ways, we have a whole chart in different ways. It is using AI and even more, more advanced like Agentix. Then the second one is Contact center, you know, and then no surprise there, that's like one of the big areas of using that uses AI. And then the third area is actually sales. So those are kind of the top three. There are others as well using it. But if you look at kind of a bar chart of the, of the most common business areas, those would be the top three. And then in terms of functions like what business function do we use AI for? Number one is customer service, number two is automating business workflows. And then the third one is workforce management. So really being able to use AI to automate things like forecasting and scheduling and you know, being able to help people work more efficiently, better with AI agents and AI assistants, for example. And then the next area would be like cybersecurity and risk management. We're really seeing a lot of uptake in AI in those areas and there are plenty of others as well in the study, but those are some of the top ones that we're seeing at this point.
[00:04:32] Speaker A: So it's everywhere, essentially. Right. So yep.
And with that in mind, then one of the things you asked about in the study was the AI center for Excellence. Right. And a lot of companies planning or, or that already have, so 38% have one right now, 58% planning. How important is this, do you think, for kind of long term overall success with AI?
[00:05:01] Speaker B: Oh, I think it's really important. I think with any sort of new, very pervasive or transformative technology, a center of excellence is almost a necessity. Especially with larger companies where you have, you know, midsize to larger companies where you have different business units and all of a sudden a new technology emerges and you see all these silos because every business unit's sort of doing their own thing and then there's no consistency and you start breaking security policies and things like. So, you know, you could think in the past you'd see companies standing up a center of excellence for like wireless collaboration, cloud, any, any sort of transformative technology. So I think it's very important for the success of AI. You want to have different viewpoints coming in which you get with the center of excellence and you want to get consistency in all of your processes and how you're regulating the technology. And if you don't have one of those like a center of excellence and regular meetings and a regular group of people from all different disciplines, you're really missing out on a lot, I think.
[00:05:59] Speaker A: So is it too early for AI? Center of excellence is just sort of true value at this point, or. I mean, what's our data showing us?
[00:06:07] Speaker B: No, no, our data is actually backing that up. And I think you already mentioned we had about 38% of companies that already have a center of excellence stood up and our research success group. And those, those are companies we ask about a series of questions that help us to identify the companies that have higher than average measurable business improvements in the metrics, various metrics that we look at. So the success group, they're kind of, they're showing all sorts of improvement in revenue, in cost savings, in customer satisfaction scores, in employee efficiency, all sorts of things like that. And when we look at our success group versus our non success group, which are those who are below average, they still may be seeing some improvement, but it's just not as good as a success group. That success group is nearly two times more likely to have an AI center of excellence than the non success groups. That's just a correlation, but it tells us that that's a common practice that successful companies are doing.
Another thing I would say is that, you know, when you look at companies that have a center of excellence, their actual business metrics improve significantly, significantly more than those who have no plans for a center of excellence or even those who are saying, you know, we're planning to do it, but we haven't done it yet. So for example, when you look at companies with a center of excellence, they're seeing a 30% increase in sales. Those with no plans are only seeing a 15% increase in sales. So great, they're still seeing an increase, but they could be even more if they had a center of excellence stood up. Same thing with like csat. With the companies that have a center of excellence, their CSAT score goes up by about 35%. Those without one, it's about 25%. So you definitely see some hard numbers that kind of back up the importance of one.
[00:07:55] Speaker A: Yeah, that's great. Love the idea. But then having the data to back it up is so important.
We've talked about the two of us and the Metro G team have talked about the trend that you saw among research participants and they're thinking about how they're going to work with technology providers, how they want to work with technology providers around AI. So why don't you share with our listeners what you discovered?
[00:08:24] Speaker B: Yeah, this is a really interesting one, especially this round of research. So about a year ago, actually when I did this study a year ago, I had a hypothesis that said, you know, AI is moving everything so quickly, the pace of innovation is going so fast that companies are not going to be able to continue buying technology in the same sort of old school IT way that they used to use. And what that was, was okay, we have an in depth understanding, or at least we think we do an in depth understanding of technology. And we need to really understand what you're doing before we're going to buy from you to solve a problem or address an opportunity. And that often meant like an IT person and a technical person at the vendor, you know, going head to head and arguing maybe and discussing how are you doing this and why are you doing it this way. And you know, getting into a lot of cycles like that before, they're like, okay, now, now we're convinced that you're doing the right thing.
So that's one way and that's the way companies have been buying. But because of how quickly things are moving now, my hypothesis is it's going to switch to, I just have to have a high level understanding of technology and I'm going to rely more on my technology partners to make recommendations to me on what I should buy to solve my problem or address my opportunity. So a year ago it was roughly 2/3, 1/3, 2/3 saying, oh no, we have, we need to have a high level or I'm sorry, we need to have an in depth understanding of technology before we make a decision. And about a third roughly saying they, they wanted to have, you know, more of a, you know, high level understanding, rely on their technology providers more and that it went down, those numbers narrowed a little bit just by about five points six months later. And then we just did it again now. So a year after that first question, the percentage that said that they wanted to have an in depth understanding went from 68% down to 51%. Those who said they wanted to have a high level understanding went from 33% up to 49%. So we're almost at 50, 50 right now in how companies are buying. What this really means is that your decision on who you're going to select as Your technology partner is more crucial than ever because you're relying on them, you're trusting them, they have to be a trusted advisor. You're going to come to them with your problems or your opportunities and trust that they're going to give you the right solution that will help, you know, address them. So very big shift and it means a lot, it really means a lot of change to the vendors themselves because when you think about it, the companies that are more apt to say we want, we want a high level understanding or relying our vendors are smaller companies, there's more of those and the smaller companies, the vendors aren't going to be able to serve all these, you know, small and lower end of mid sized companies that you know, want to have their, they just want to rely on them, they can't keep up, they can't keep up with the pace of innovation. So I think this is going to mean a lot for channel partners. It's going to tell us that these channel partners are going to have to do a lot more consultative selling. Maybe they'll be charging for that and you know, new opportunities there. But definitely a shift in how companies are going to be buying.
[00:11:30] Speaker A: Well, we'll have to see what, what happens what a few months from now when you do your, your CX study, where they, where we land with this.
[00:11:39] Speaker B: Yeah.
[00:11:39] Speaker A: Okay, so we talked about some operational strategic things. Let's talk about technology now. So in this study, this is the first time that we've asked specifically about agentic AI which that term has just become so buzzy in the last six months or so. So what did you ask? What did you learn about company's approach to Ejectech AI today?
[00:12:02] Speaker B: Yeah, so very interesting. Only about 52% of companies are even familiar with the term.
So you know, we've got a long way to go from a technology vendor standpoint to make sure that, you know, people actually know what this actually means.
[00:12:21] Speaker A: What does it mean? Oh yeah, yeah, right.
[00:12:25] Speaker B: As you can see, you know, we have a Doberman in the background now, so that's always trying to get him out of the way here.
Yeah, so we, we actually did and that's a good question. We asked companies what, what do you, what, how do you define agentix? And there were about four different common answers that we saw. Our definition of it is that it's an advanced AI framework that uses large language models, so part of generative AI to make decisions and take actions aut autonomously. So in other words without human involvement at all. So in action agentics basically Help they help humans in the moment or they automate an interaction or a process.
So I think that the autonomous nature of agentics is a really big piece of it. So I look at the word, it's kind of the combination of two words, agents and analytics. And so when we're looking at the LLMs, LLMs, the LLMs are analyzing what's going on and taking action autonomously based on what's going on. Now I think moving forward, when we look at like Agentix 2.0, that's going to expand to include like a hierarchy of AI agents. We have like the master agents, the sub agents, the assistants, all these bots working together to analyze an issue and take action and close the loop. I, some will say that there are vendors that do this already today. I don't think we're quite there today. I think we're still early, very early stages of Agentix. But yeah, we don't see a lot of companies familiar with the term. Among those who are they still at this point find more value with like functional task specific prompts compared to agentic AI prompts where it's real adaptive decision making. Actually I don't even need prompts. It's agentic AI for adaptive decision making and they're finding that functional tasks specific prompt more valuable by about 2 to 1 right now. So we're still very early stages. I don't think a lot of companies really see what the value is yet or have experienced the value. Some have, but not, not very many. When we ask what their plans are with agentics. Most are going to use agentic agents to complement existing AI agents versus like a rip and replace type of thing. I think over time you'll see some shifting going on there. But when we look at all the different types of AI that companies are using or planning to and spend Money on in 2025, the biggest growth area right now is Agentix compared to every other area we looked at.
[00:14:58] Speaker A: So the evolution of generative AI, more or less, I mean roughly.
Okay, you mentioned prompts a couple times. What about prompt engineering? That was another.
Yeah, that was another good area for the study.
[00:15:13] Speaker B: You know, as a practice when we think about prompt engineering, so that, that's basically crafting a prompt to put into a large language model in such a way that the output gives you exactly what you're looking for. And there is a science to that. In practice it is vital to 58% of companies. So that's, that's very high for vital because the next line is important. So you look at the difference between vital and important. I mean, that's huge. 58% of companies saying it's vital. Just coincidentally, the same percentage, 58%, train their employees on how to engineer prompts. So they're teaching them. Here's, here's what you need to say. Here are the, here are the details you need to provide. If you're asking this type of question, this is what you need to specifically tell the large language model to do.
But the other thing is a majority are also interested in prompt templates.
So having a provider give you a list of templates that either you can use to manually enter or you have as kind of like in the middle, I'm entering something, just a natural language as a, as a consumer, as a, you know, as a human, I'm just entering something in natural language. And then before the large language model actually takes it and processes it, there's something like in the middle, basically that's looking at what I entered and what do I really mean by that and clarifying and fine tuning that prompt before it actually goes into the large language model to respond. And so these prompts, either way that they're being used, are really in demand right now. The top value is, is that they get quick answers and the questions are asked properly to AI. So I'm going to get what I need and I'm going to get it quickly. And the way it's asked into the large language model is done in a proper way. The problem, one of the reasons why companies aren't using these right now, by the way, is that their provider doesn't offer them. So that's something for vendors to be thinking about. We have a lot more in this area, a lot more questions we asked around prompt engineering.
But this gives, this gives you a little bit of an overview. Obviously, clients have access to all of the charts. There's about 200, you know, just baseline charts in the study. And we also do all sorts of cuts by vertical industry, by size of company, by global region, all sorts of data here. But it's, it's a big study and there's a lot more to prompt engineering. We'll be doing some reports on that coming up as well.
[00:17:39] Speaker A: Very meaty study.
Absolutely. A lot to, lots to think about in this study.
But let's, let's go through a few more points here. So what if I'm just getting started, I, you know, I, I haven't introduced AI for my, my customer experience operations, my contact center at all.
What should I be considering? Which AI technology should I be considering if I have, you know, specific business goals, whether it's increasing revenue or improving CSAT or customer ratings in general.
Have you found kind of tracked which, which technologies trigger which metric the best?
[00:18:17] Speaker B: Yeah, we've done really detailed analysis on this. There's a lot, this goes pretty deep. So if this is something you're interested in, we've done, we've looked at cuts by our research success group, we've looked at it by verticals, by sizes of companies. So there's a lot there. I can give you some, you know, kind of high level. We basically evaluate the percentage improvement of specific AI technologies when it comes to certain business metrics. So we basically ask, okay, if you're using AI voice agents, how has that affected your revenue, your customer satisfaction, you know, your, your costs, your agent efficiency? And then we go down the list of all different types of AI. So as an example, like if you want to increase revenue at your company and you want to, you know, you have budget for AI and you want to know, okay, I can't do everything, I can't boil the ocean, nor should you, by the way, what are the top technologies that are going to help the most with revenue improve? Improvement. And in our research, AI powered content creation actually is number one, followed by Agentix. So having those agents be able to, you know, AI agents be able to act independently, act autonomously, translation. So being able to expand into new countries, new regions without having to hire people with the same language who speak the same language, inferred sentiment. So that's just basically telling you what are our customers, you know, what do they think of the interaction they just had even though they didn't respond to our survey, Can I see what we're doing right and wrong. So we can continuously improve and then AI chat agents. So being able to have that capability to interact in real time with natural language in a self service channel where I'm using a chat agent, those are the, those are the top and there's a long list, but those are the top ones that actually moved the needle the highest on revenue. Now don't take that to mean that's the biggest adoption because as we already said, Agentix is a low adoption so far. But those who are adopting it are seeing good results. So kind of take it that way. But yeah, we've evaluated a long list of AI technologies and how they affect the key business metrics. So if that's something anyone is interested in, please reach out to us and we'd be happy to help you with that.
[00:20:28] Speaker A: Okay.
So a lot of great information you've already shared. Let's wrap up.
Beyond what we already talked about, what are some of the other most interesting or most exciting things that you found in the study?
[00:20:43] Speaker B: This is a loaded question because there's so many. And when I was at, we were recently at Enterprise Connect talking to a lot of companies about this and they were asking me that same question. I was like, oh, where do I start? There's a long list, but I'll point out a few. And we're going to be doing a lot more, you know, like webinars and content reports and things like that on this data. There's a lot there. One of the things I would say that's that I thought was interesting anyway was that most companies are seeing an ROI right now with their AI deployment. And when we ask them what's a reasonable wait time to see an ROI? You know, with technology it's usually like 18 to 24 months. You know, with ROI for AI, it was definitely a bell curve. The tip of that bell curve was 6 to 12 months. And when we looked at our research success group, less than six months. So you know, technology members don't have a lot of time to prove their case to show an roi. So that was one of the big ones. I think. The other thing, you know, we talked earlier about how every company is using AI in some way, shape or form, but only 27% have a single executive in charge of like company wide AI strategy today. So that's concerning to me now a lot say they're planning to do this, but that's concerning to me because now we have silos and a lot of different problems because of those silos and then revamping and ripping and replacing and oh, we've got to get everything consistent now. So I would say if you're, if you don't have somebody in charge of AI, that's something that needs to be done right now. The most common title we see there is like cio, cto. But there are a lot of new titles also emerging with AI in them, like Chief AI Officer, head of AI. So that's actually emerging as a title.
Another kind of fun and kind of quirky and interesting1 is 68% of companies say that AI can make anyone a salesperson. Just through that real time coaching, it can make you a salesperson.
You know, we always hear about AI cutting jobs. Oh, AI is going to, I'm going to get replaced by AI. But here's a real interesting one. More companies, higher percentage of Companies are actually adding new positions and eliminating them. So I think that's a, you know, pretty big aha moment, right? We talk, there's more data scientists being hired, more data analysts, more prompt engineers. Like there's all sorts of. We have a whole list of positions that are being added, you know, to the, you know, to companies now and a couple more I would say. I'm really into conversation Interact and Interaction analytics right now. I think LLMs can go into every conversation either listening, using speech analytics or looking at transcripts and pull out a ton of like nuggets there. And right now a lot of that's being used within the contact center. But I had a hypothesis that, that these should be used outside the contact center in like executive level dashboards. You know, pull out the information that the executives would want, you know, might not be as detailed, it might be more high level. But 84% of companies agree with that. They think that Interaction analytics should be part of an executive level dashboard. So that's another area that I think, think, you know, vendors can be looking at to, you know, potentially provide to their customers and lots more like I said. And we'll be talking about that in upcoming, you know, webinars and reports.
[00:24:07] Speaker A: Okay, so speaking of upcoming things, what is next on the CX agenda for us?
[00:24:16] Speaker B: Oh gosh, a lot more studies. So we have our CX Metrocast, our annual study coming out here in April that's looking at all sorts of adoption numbers and oh, are you leaving your provider? Why? If you're leaving your provider, who are you considering? You know we have of course always our quarterly updates and you know, just looking at rankings and Metro Stars, which are our customer. Customers rate their providers on a variety of different areas of sentiment and business success. That'll be coming out in April as well, I believe. April, late April.
We also are in the field with a workforce engagement management study which is going to be awesome. Really good one there. We have a knowledge management study. Bet that you are going to be really, I'm really excited about that one. Just data and knowledge management. Like you know, we talk about AI all the time but if you don't have that strong data foundation that could be a big issue. So that's really going to be a cool study we do. Our annual customer experience optimization study will be coming out so lots there. We do have also something we're calling the Expert Insights program. We have four different topics we're going to be doing in 2025, 2025 in the area of CX. A lot of AI in there. Like, we're going to do one on Agentix, you know, for example, we're going to do one on interaction analytics. There's we're going to do one on agents. What's going on with agents and how are we using technology to help them. We're also going to be doing one on what do consumers think and about the technology they're getting and how does that compare to what businesses think they're delivering. A lot of disconnects and gaps there. So we have that kind of program. We present our research and then up to three technology providers sponsor these webinar series and they provide really good insights too. They always have really great ideas of how you can address some of the issues that might arise. So those are some of the things we have coming up. But always lots what's going on.
[00:26:15] Speaker A: Absolutely. So a ton of, a ton of great stuff in AI for Business Success. Love to get you guys briefed, so reach out to us. And lots of stuff coming down the pike for CX.
In the meantime, be sure to visit
[email protected] check out that AI for business success study. It's up on the website. All of Metrogy's leadership on AI and CX and beyond. And as always, we're really happy to hear from you. So feel free to reach out to us. You can use the contact button on the metrogy website or just reach out to either of us directly via email, robinetergy.com or bethetragy.com and with that, that's all for now. On behalf of Robin and all of the Metrogy team, goodbye till next time and take care everybody.
[00:27:02] Speaker B: Thank you.
