Customerland
Customerland is a podcast about …. Customers. How to get more of them. How to keep them. What makes them tick. We talk to the experts, the technologies and occasionally, actual people – you know, customers – to find out what they’re all about.So if you’re a CX pro, a loyalty marketer, a brand owner, an agency planner … if you’re a CRM & personalization geek, if you’re a customer service / CSAT / NPS nerd – you finally have a home.
Customerland
How TP USA is Transforming Customer Experience with AI
The relationship between artificial intelligence and human emotional intelligence isn't just fascinating - it's reshaping how global enterprises deliver customer experiences. In this revealing conversation with TP USA CEO Mike Lytle, we explore how the world's largest customer experience provider is thriving in the age of AI by finding the perfect balance between technological advancement and human connection.
Mike shares TP USA's remarkable evolution from traditional call center origins to a diversified business where specialized services like trust and safety, health advocacy, and visa processing now represent half of their operations. With nearly 500,000 employees serving clients across 170 markets in 300 languages, their global perspective offers unique insights into what's working at the intersection of AI and CX.
What sets TP USA apart is their philosophy of being "powered by emotional intelligence, enabled by artificial intelligence." As Mike explains, developing emotional intelligence at scale requires sophisticated approaches like their TP Interact tool, which analyzes customer-agent sentiment and provides targeted coaching opportunities. This focus on human capabilities becomes increasingly important as AI handles routine interactions and agents tackle more complex, emotionally charged situations.
The conversation offers practical wisdom about implementing AI successfully: start with specific use cases rather than company-wide transformations, recognize that AI implementation requires robust data governance, and build momentum through early wins. Mike reveals that over 700 TP USA clients have implemented at least one AI solution, with some achieving 80% digital containment rates and 20-30% efficiency gains.
Perhaps most valuable are the insights into how AI is transforming agent enablement through simulation training, generative AI knowledge management, and emerging agentic solutions that can execute complex multi-system processes. These innovations directly address the challenge that most agent turnover happens in the first 90 days when employees don't feel adequately prepared.
For CX leaders navigating the AI revolution, this conversation offers both strategic guidance and tactical examples of how human-AI collaboration can create better experiences for customers and employees alike. The evidence? TP USA's strong financial performance despite predictions that AI would make traditional customer operations obsolete.
When you start an AI journey, you're also starting a data journey. Your data management, governance, cleansing all of that work is relatively significant when you think about it as a whole company AI transformation and so going use case by use case you have a much higher probability of success and the savings or benefits that you collect from those early implementations can go on to fuel more and more.
Speaker 2:Today on Customer Land, mike Lytle, who is CEO at TPUSA. You may know TPUSA as teleperformance, but it's a bigger company with a bigger reach than I think many people, even in this space, are aware of. So first, mike, thanks for joining me. I really appreciate this.
Speaker 1:Yeah, absolute pleasure. Looking forward to the conversation today.
Speaker 2:So maybe we can just start with what is now TP, what used to be Teller Performance, and where that company came from was up until fairly recently, and where you're going next.
Speaker 1:Yeah, tp has been around forever. We were founded in 1978. And I think a lot of your listeners probably think of us as the CX company and large global scale. We're nearly half a million people, we serve 170 different markets, 300 different languages.
Speaker 1:I think a common misconception about TP is that we're a customer care organization and certainly that's our heritage, that's where we came from. But as we sit here today talking, we do about half of our business in customer care and about half of our business in other specialized kinds. About half of our business in other specialized kinds of services whether that's our specialized team that do interpretation services, localization, visa processing, health advocacy, our trust and safety business, which has really been booming over the last five, seven years, we'll call it and the full gamut across back office services and things like that. And what I think is interesting when you get that kind of diversification in the broader BPO or BPS space is the ability to see end-to-end across an organization for places where there's waste, there's friction, there's things that just aren't working. Quite right that you can help bring a blend of human and technology to solve those problems.
Speaker 2:I guess with a company of the size and scale of TP it was almost inevitable that you would be well, first of all you'd have a view to those kinds of gaps and opportunities, but second, that you would kind of have the ability to come in and kind of solve for some of those problems. So, strategically, I think it makes an awful lot of sense, of course. But I'm wondering, just to set a little further context for this, maybe give us a little bit about your own personal journey to this point. You know, how did you, where did you come from? How did you get here?
Speaker 1:You know, how did you? Where did you come from? How did you get here? Yeah, it's one I've told a few times because I've been in this business for 26 years and at TP for the same.
Speaker 1:Honestly, I was in school and I wanted a part-time job so that I could get a car, and never had any aspirations to make a career in the call center.
Speaker 1:But like many people who enter a call center, they find a way to fall in love with it and for me it was first about helping people within the call center be the best they could be. I had the opportunity to be a supervisor to teach people how to sell, and I got hooked instantly on that coaching and development and helping people elevate. And as I moved on through my career, I've had the opportunity to live and work all across the US, mexico, philippines, china, malaysia and on and on. So that's created huge life experience for me that I don't think I could get in many other organizations or industries. And as I've grown up through the organization, while there's still a huge amount that I take satisfaction in helping people develop their careers in equal measure helping our clients solve their problems, because across the 1,500 clients that we have at Teleperformance, they each have their unique problems, opportunities, things they're trying to solve, for that always keeps you engaged and interested in what's coming next.
Speaker 2:I can only imagine If we could just take a half step backwards, because as we were introducing ourselves here, you kind of ran through a list, a menu, if you will, of the services that TP now offers. That weren't really part of what TP is originally known for and I kind of want to just unpack a few of those. But specifically the trust and safety operation that you mentioned I was doing a little bit of research on and you know trust and safety is a very high level kind of a phrase. It can mean a lot of different things, but maybe tell us a little about what it means in the TP world and how that's kind of worked out and in the market space.
Speaker 1:Yeah, trust and safety has been an interesting one. I think if we were sitting here talking 10 years ago, we wouldn't even have the term trust and safety, and initially everybody thought of trust and safety as content moderation, and while that's a huge part of what is in trust and safety, there are things that we do in like know your customer, fraud, detection and prevention, financial crimes that all sit under that trust and safety umbrella, and really trying to find ways to leverage that human capability to sort through information. And I realize in today's age, most listeners are probably thinking well, that's what AI is for. We all know that AI has a certain level of confidence and ability to determine the right outcome, but there's always an error or an exception or a component where you want to keep the human in the loop, to continue to tune the model to become more accurate and more precise and that's really where we see TP come forward is in that last human element and tuning the model to make sure that it's fit for purpose.
Speaker 2:Interesting. I was super intrigued as I was kind of going through some of the materials that your team had sent over at the way. Tp is kind of framing up the company narrative and I know there are several components to that, but one in particular, powered by emotional intelligence, enabled by artificial intelligence, and you can't see it because it's an audio medium, but it's a very cool graphic. That caught my eye there. But I speak to a lot of people who operate in the CX slash, bpo, bps space and I think those kinds of ideas are kind of out there in the ether. People acknowledge them but really nobody really puts it in ink. You know, you put that in writing all of a sudden. That really means something different. So you know, would you say that's I was going to? I was going to ask is that a core value for TP? But that's not even fair because you know, five years ago that wasn't really even a thing. You know, yeah, so where does that come from?
Speaker 1:I think the emotional intelligence is innate in people, right, we all have a capability or capacity for emotional intelligence, and what's unique is distilling the attributes so that they can be coached and developed at scale. And that's something that I would say has evolved over the last handful of years, and I think it's out of necessity in running a large scale people-based operation as automation takes hold and interactions become more complex. Complex interactions are more emotional by their very nature. Right, they're challenging. We don't have a clear resolution. It's something that we have to be collaborative and, as I see our frontline agents going through this, what I talk to our leadership team about a lot is we often mistake empathy and sympathy. Sympathy is very easy, it's easy to say sorry and those kinds of things, but to truly put yourself in the other person's shoes, to feel what they feel in that scenario and to connect and communicate with that level of empathy is exhausting. So being able to build emotional resilience is equally important to the ability to develop the emotional intelligence itself.
Speaker 2:Um, I think you mentioned a couple of things that are worth really spending some time on there. One is that that, uh, empathy just in and of itself is difficult. It requires something of the of of the human there. That's often really uncomfortable and you really have to press forward into into that space. On the other hand, it's infinitely more difficult to try and work that out at scale. Infinitely more difficult to try and work that out at scale, which which I think is something I look, I'm really interested in the secret sauce that TP has to be able to do that at scale, cause you are the largest in the space and yet this is clearly something you believe in. So how do you work that out against? You know, through thousands and thousands of people.
Speaker 1:Yeah, it's a constant challenge and opportunity. I'll call it that way. Just like anything that you're trying to develop people to be good at over time, it takes repetition, iteration, to come at it from different angles and what's the operations is always going to be a little bit different. So to bring the scale element, we have a tool called TP Interact and in most organizations you probably think of it as interaction analytics which is the AI version of our historical QA.
Speaker 1:Within that tool we can look at the sentiment both on the customer and the employee side, to understand how they're evolving through a conversation.
Speaker 1:We can then look at that down to an interaction level and understand which types of interactions they're having with customers, where they're the best suited from an emotional intelligence and from a resolution perspective, and what you'll find is those two things are often linked the more equipped I am to resolve, the easier it's going to be for me to empathize, because I'm focused on the customer rather than focused on trying to navigate that resolution. So we bring that information together into a coaching dashboard for our frontline leadership to help match the tactical components of emotional communication and the tactical hard components of first call resolution. It's something that, again, we have to continue to develop and drive every single day and truly, you asked about the secret sauce. That is it. Tp operates with TOPS. It's our core operating process and we fully integrated the elements of AI and EI into that day-to-day coaching and development process no-transcript and what a great moment we all seem to be living in right here.
Speaker 2:But I feel like many of us and I'm including myself in this space are still at this kind of moment where it's this mesmerizing shiny object that does some really neat and powerful things, but I just don't feel like we are fully considering the power of what AI can do to enrich human interactions. And I wanted to just kind of, if I may, just talk about and get your thoughts on how you see AI being deployed and leveraged in your organization for specific purposes and where you think TP may be going with that whole idea.
Speaker 1:Yeah, it's a huge question, like you said, and there are a lot of different routes that you can take. I always find it's helpful to center on the fact that we in the CX domain have been through multiple iterations of technology that was meant to revolutionize what we do and, in some cases, eradicate what we do, depending on who you'd ask.
Speaker 1:Whether it was the initial IVR that was going to take huge amount of volume out. It was the online self-help, it was the mobile app, then it was the chat box, the conversational chat box that we've all come to know and love or have some other feelings about. I think it's important to put that in context. We always think that these technologies are going to have massive impact and over the long term, they do fundamentally change how we operate our business, but we incorporate them. They enhance the interaction that we have with our customer and we make resolution more accessible to them. So when I think about AI, I try and think about it as the multiple iterations that we've already been through. If you rewind back, even 20 years ago, companies were already using RPA, which is a low level of artificial intelligence, to automate simple processes, to do logins on behalf of agents, simple things like that. We then moved into the conversational AI domain and we started to program interactions and get a certain level of proficiency and outcome from a decision flow, a workflow that's presented to a customer into generative AI, and everybody was taken aback by chat, gpt and the conversational nature that was previously unthought of. Within an automation tool Leveraging generative AI, we can have a much more natural conversation, but it's still limited in part in its ability to take action, to do all of the things that an agent can do across multiple systems. But it's great at FAQ, it's great at knowledge management, it's great at call summary things that are highly repeatable and take a lot of time for an agent or a customer to be able to process on their own. So I think there are some great use cases and we can dive into that.
Speaker 1:The one that everybody's talking about today, agentic AI, where the AI can actually take actions within the systems is one that's hugely exciting but also scary for a lot of organizations. When you think about a fully autonomous AI agent that's doing things within your system, most companies are still thinking they want some human in the loop. Most companies are still thinking they want some human in the loop, they want some supervision, they want some control framework around that, and I think that'll be important for the foreseeable future as organizations mature in their AI infrastructure, because if you think about all of those different AI layers that we talked about, it's not as if you retire your RPA to go to conversational or retire your conversational to go to Gen AI and on to agentic. You have a blend of these different AI solutions within operations that need to be managed, maintained and enhanced as the capabilities of those tools improve. So I know I said a lot and I know you want to probably dive into a couple of key examples, but that's my thought on AI at a high level.
Speaker 2:I think it will be again. What I think just happened here is we prompted a second version of this conversation somewhere down the line so that we can really unpack some of that, some of that, but I think, just for today, understanding how TP is viewing AI at a high level and how you're deploying it is really instructive. I mean, you're going to find that there are a lot of CX professionals that are listening to this, that are paying attention by virtue of the fact that I'm talking to the uh, the uh CEO at TP USA here on the podcast like this is worth listening to. So I think your, your opinion and views carry some weight.
Speaker 2:Um, that being said, even in this industry and I you know there's a, I think, first of all, it's an industry of a lot of smart and dedicated people, but AI still carries with it a certain amount of call it fear and trepidation. It really still does require human touch to really refine it and make it truly empathic empathic, but because of the number of companies that you deal with worldwide, do you see any kind of commonalities to the I would just call it AI maturity level? That is kind of moving certain companies further along the curve versus some that are lagging. I mean, are there things that CX operators or maybe the people who run these companies themselves, can be thinking about to kind of either catch up to the curve or accelerate themselves along it, just based on things you see in the field?
Speaker 1:Yeah, and if you don't mind, I'll answer your question in my language a little bit.
Speaker 2:When I think about the.
Speaker 1:AI maturity of organizations. I think about it in the context of how far they've moved automation, self-help, been able to provide resolution in the customer's own hands prior to getting to a TP expert. The industries that I see that are furthest along that automation and self-help journey are typically the travel and hospitality and telco industries. On the highly automated self-service side of the spectrum, and on the less automated and self-help, you'll find things like utilities. It sounds weird to say banks, because we all have our mobile apps and we get to do a lot of things there, but there's still a huge amount that is face-to-face and phone-based within banks and, of course, in the healthcare system. Those skew toward the less mature side of automation, at least from a customer or patient interaction standpoint. Your second half of that question about how they can accelerate it really boils down to two things, in my opinion is one finding the use cases within your organization that are going to provide the most benefit to your employees, your customers and, ultimately, the organization, and using that as a testbed to learn how to implement AI.
Speaker 1:Well, because AI, a lot of people think, because ChatGPT and others are so easy to use that I'm just going to plug in an AI model and it's going to learn by itself and it's going to take care of all these issues in my organization. But when you start an AI journey, you're also starting a data journey. Your data management, governance, cleansing all of that work is relatively significant when you think about it as a whole company AI transformation and so going use case by use case, you have a much higher probability of success and the savings or benefits that you collect from those early implementations can go on to fuel more and more and, just like every organization, as something delivers a result, we tend to put more resources into it and you can start to build a bit of an ecosystem, and whether that's through partners like TP, whether that's something you bring up on your own or some hybrid model, there are lots of ways that you can leverage the skills and capabilities that exist to accelerate that journey.
Speaker 2:You mentioned a couple of things there that, again, I want to unpack all of it here, but we're just not going to be able to do that today. But one of them, I think, is one of these critically important but soft skills and it would seem to me that TP has probably mastered it, if I may and that is introducing this kind of digital transformation it's an old term now, isn't it? You know, nobody even says that anymore and only transformation Into corporate cultures that may not be ready for it or think they're ready for it, but you know there's a maturity there that you know kind of needs to be accelerated somehow, that you know kind of needs to be accelerated somehow. So, and I and I think, talking to other people in the space who are implementing technologies or transformations, it's one of those things where where everybody kind of acknowledges it and like, yeah, it's, it's key and critical to our success, um, but everybody wants to talk about their product or their service, and I think so many people who are listening to this and read Customerland are way more interested in like, yeah, but that's the hurdle every one of us deals with, every single one of us.
Speaker 2:Maybe TP has an answer. So how do you do that? You go into an organization. You know you look for my words, not yours the low hanging fruit, the easy opportunities to try and build some wins. But culture is a different, is a difficult thing to move the biggest barrier to transformation initiatives, at least that I've seen.
Speaker 1:There are entrenched ways of working across many organizations and while you can put data on a page that says this is an easy decision, you should change this process. You should implement this technology. As you rightly point out, within the organization that process may have a certain reason for being. You may have a stakeholder that has a fear of change, somebody who's newly promoted that doesn't want to be the person who makes the wrong decision. There are all these internal dynamics that have to be considered to make a transformation initiative really get traction, really get traction.
Speaker 1:And I think at TP the fundamental is we don't ever try and sell technology, a product, a service, a solution. We try and engage with our clients about what are the problems that the organization is facing and how can TP bring expertise from our global network of people and technology partners to bear to solve that problem. And when I think about change management, we manage 1,500 clients globally and literally thousands of lines of business within those particular client organizations, and the amount of change initiatives that we're asked to facilitate across our half a million people is immense. So it's not really a secret sauce, I have to say. It's just a consequence of the business that we run that we have to be really good at change management. So navigating the initial decision is one piece of it, and then the second is reworking the way things are done within an organization. With that new technology layer implemented, that's where we start to help organizations think about their business differently, change the way the work is done and ultimately create some really impactful outcomes.
Speaker 2:Yeah, that's way different than just being a BPO, a large BPO. There's a lot more that goes into that. That in fact I think it would be great fun you know again two or three more conversations down the line somewhere and talk about the evolution within TP of that skill set. Skills are very kind of human empathy based with a, with an awful lot of corporate dynamic expertise kind of baked in there, and I don't think you kind of just wake up one day and go. You know what we really need, a change management kind of function here. Let's do that for a living, Cause it's just, it's wildly complex looking at all the different you know business lines, aspects, cultures, silos, kpis and everything else that could be standing in your way.
Speaker 1:so yeah, you said it's. It's an interesting challenge to tackle um, and one that we continue to to evolve um, I want to jump to a news item.
Speaker 2:I I was doing a little bit of research ahead of this call and I'm just going to read the headline, because it's easier that way. From late March of this year, Teleperformance, the world's largest call center operator, has posted last month's strong full-year 2024 results, surpassing analysts' forecasts of revenue of 10.28 billion euro and a recurring EBITDA margin of 15%, despite concerns that artificial intelligence might render traditional customer experience operations obsolete. I think I read that and I'm like I want to post this as a headline across our website because everybody in the space is going. You know we got to track and push with AI, but ultimately we're putting ourselves out of business, and yet TP has proven otherwise.
Speaker 1:Yeah, I mean. First, thank you for the compliment and the numbers we were, of course, very proud of. I think you know. From my perspective it's twofold.
Speaker 1:One we have seen good progress in our AI initiatives.
Speaker 1:We have over 700 of our clients, which is about half, if you're doing the math, of our clients who have at least one AI implementation with TP, and typically when we drive good outcomes for our customers, they reward us with increased share of wallet or new opportunities, new challenges, new lines of business, new things to take that benefit and apply it to.
Speaker 1:And when I say largely impactful, we have AI implementations that increase efficiency or productivity. 20, 30%. We have individual interactions because, again, the way we like to look at it is per use case, where a particular call type can see an 80% containment rate in a digital channel where previously those were coming through to TP experts. So those kinds of benefits, as you'd expect, can get clients excited about what's next for them in AI and how to do more work with a company like TP. The second side of the equation as we started our conversation, customer care has been a huge part of our history, our legacy, but all of these other lines of business within the broader BPS space are growing at an even faster rate than the core customer care work and something that we continue to invest and grow, so I think that dual engine really provides a great growth trajectory for TP.
Speaker 2:It's a model worth paying attention to. I mean again, I'll just reiterate I think the CX space is populated with some of the smartest, most energetic people in business, always looking for a way to solve problems in efficient and effective ways. But I think taking a good, hard, high-level look at the way TP has built your models they're all synergistic and all kind of play to and support the strengths of the kind of core business is really smart. I think there's a lot of great takeaways from that. Maybe that's the fifth episode of this conversation. We've talked about call it 270 degrees around this sphere here. But we've missed one giant chunk and that is agent enablement, and I know that's a big part of what TP is focused on. Can we talk about that a little bit?
Speaker 1:Absolutely I'd love to. As somebody who started out as an agent on the phones, I have a giant soft spot in my heart for our frontline teams and appreciation for how hard the work is that we ask them to do. So anything that makes them more ready for the job, that makes that job easier, allows them to be there for their customer, is a huge step forward for me when I think about how we deliver services. If I rewind back the way we used to do, training used to be facilitator-led and then we had this aha moment where you know what actually we should probably have the content match the work that's going to be done and we started redesigning curriculums and thinking we were really great and we saw great outcomes Fast forward.
Speaker 1:Today, with AI, we can do a simulated training environment where we look at the exact workflows that an employee needs to do when they actually interact with the customer, that an employee needs to do when they actually interact with the customer. The AI will monitor how they're navigating those screens, that they've done the right workflow and they get a score back on how effectively they understand that particular work that they're doing. We now have the ability to build AI agents on the other end of that, so they can actually practice their conversation around a particular interaction type with an AI agent and, again, with interaction analytics, get a score that says I now am proficient at navigating the system and having a conversation with the customer, before I ever graduate from training and talk to a live human about a particular issue type. That, to me, was a phenomenal revolution in the way training is done, and one that not every client is yet taking advantage of.
Speaker 2:That's really really interesting. As you can imagine, everybody's kind of dealing with retention and enablement on some level in this space. I don't think I've heard of anybody really approaching it with that level of sophistication. Maybe that's due to your size of scope, Maybe not everybody can afford that, but that's a really cool way of approaching this.
Speaker 1:Yeah, and I mean, if you look at our business, the vast majority of employee turnover happens in the first 90 days because they don't feel ready for the job. And this is a great way to reduce that burden on the TP expert while improving the customer outcome. So the ultimate win-win kind of a scenario, brilliant.
Speaker 2:I think Mike go ahead.
Speaker 1:I can throw a couple others out there if you want. I just wanted to give you a chance to bounce those back.
Speaker 2:I think you know I'm here on your schedule, so if you want to continue, I am all about it.
Speaker 1:Okay, great. Another AI implementation that's near and dear to my heart in agent enablement is knowledge management. If you talk to frontline agents, most agents don't like their client's knowledge management tool because what typically happens? They go search for something, they get a list of links back. They get articles that they read all the time. Their customer is on hold waiting for this information. By ingesting a client's knowledge management tool into a Gen AI engine, a frontline employee can ask that GenAI tool a simple question and get a simple answer, and that seems straightforward. But still many clients are using that traditional knowledge management system and missing out on huge opportunity to take stress out of an employee's day to improve average handle times and make significant steps forward in customer experience.
Speaker 2:Um, you're kind of prompting a couple of thoughts there, but, um, you know, if the, if the, the next iteration of, of agent enablement is to to use and, uh, you know, enhance their efficacy through agentic AI, what that that could be doing through agentic AI and what that could be doing, one, I think makes a ton of sense. But two, what do you think that might look like in the agent's world, the agentic agent?
Speaker 1:Yeah. So the way I think about it, or the easiest way to explain it from my perspective, is if anybody's built a macro in Excel or if they've built a simple automation that helps log in or manage different systems. I think about it that same way from an agentic, real-time solution. A lot of our customers still have multiple systems that a TP expert is using in their interactions with a customer. So I may have my main screen that I'm interacting with you on, but then I have five, six, 10 different screens that I'll have to use to execute different parts of the workflow.
Speaker 1:An agentic solution could be built for, let's say that we're having a sales conversation and I want to process an order for you. In many scenarios that's a five to 12 minute process for me to enter all the information in the different systems and make sure that's all done. Meanwhile I'm trying to keep some pleasant conversation going with you. To cover that time, an agentic solution could be built to execute all of that data entry, all of that quality checking, in one simple keystroke, versus having to do that all manually. We also could then execute an IVR-based T's and C's, because that's one of those scenarios where you're asking the TP expert to be a robot, when what you really should do is ask the robot to be the robot. And so those are some simple ways to bring forward agentic solutions into a real interaction and not overcomplicate the world.
Speaker 2:Just to make things easier. Yeah, how far down the line do you think that is?
Speaker 1:The technology is there, it's capable. It's a matter of that use case development and, importantly, ensuring that you have somebody who's monitoring your AI use cases to ensure they stay relevant and in control, because you could imagine, if you use that use case of sales order entry and you change the field and your process now breaks, you have a wild impact to your operations in terms of being able to handle interactions, revenue generation, all of that and you mentioned earlier. Companies are apprehensive about that, I think, because they haven't built that AI governance that makes sure that their agentic AI solutions stay fit for purpose.
Speaker 2:I was talking to a company just recently it might have been, in fact, it was earlier this week that deep in the AI space they develop AI applications and models for all kinds of other companies and they've got their own products out there. They've got a really, I'm going to say, cool, agentic AI solution that they actually they said we could bring it to market right away. I mean, it does amazing things. It will absolutely produce the aha moment to anybody who's using it.
Speaker 2:Said but the hardest thing we've had to do was hold off for quality reasons. We have to make sure this model is trained so accurately that nothing can go wrong, because in that particular case case, if anything went wrong, everything went wrong. So they had to kind of take the bitter pill of like, look, you know the, the publicity that they could generate, getting out ahead of the curve, ahead of the crowd, all those things, those, those real world tangible benefits. They had to say no, we, we, we just can't. We have to make this essentially a perfect product. So, um, it's. It's a funny kind of kind of tension in the. The AI world right now is everybody's moving so fast. Yet, you know, I think the smarter plays are moving really fast, but also being really cautious at the same time.
Speaker 1:Yeah, I agree and I feel for organizations that are trying to navigate this super exciting time, because, I mean, my stat says that in 2024, in the US there were 5,509 AI startups that received funding. If you think about trying to sort through that immense number of potential partners, all with super interesting and compelling value propositions, it's a huge undertaking for any organization to do. And then you move on to actually implementing, running, maintaining these systems across large organizations. It's a huge thing to undertake and I think it's something that is not likely to slow, at least in the next handful of years.
