This is why most AI projects fail

Thanks, La Look forward to it. Look forward to our discussion.

Sure, so Ascent Business Partners was actually a spin-off from a venture capital firm out of Palo Alto, Carria Venture Partners. And What Caria very uniquely realized literally over four years ago, which is early, was that they wanted to invest in AI companies, early stage AI companies. and but was very unique about this venture capital firm was that it was investing in AI companies in the contact center space. So I I I got familiar with them and the issue was that they were investing in a company. You have a twenty-six-year-old PhD from Stanford, comes up with a great company. What's his or her biggest challenge? Go to market. So Ascent became the go-to-market for the Caria investments. And there were about six or seven portfolio companies that they were helping market. fast forward about two and a half years ago I came on board and I eventually took out all the investors. turned it into an employee-owned company. and what was more important to me was yes, it was great that we were representing investments for Karia, but I've always been a technology agnostic approach. So in even in my previous company, which you know we were fortunate to sell the Tech Mahindra we wanted to give customers more choices, not just the investments. So over the last two and a half years now, we have built up this network of AI partners that can solve just about any issue in the contact center industry. we we now have over 50 partners. So now anytime we go to a customer and they define a use case for us or our business challenge. we can solve that and give them multiple options for them to look at, not just one client or one company, but that's really kind of the premise of of really kind of how Ascent was formed.

Yeah, so I mean the reality with AI is that, you know, the last two or three years has been quite interesting. In twenty twenty four, if you were adopting AI, you were an early adopter. you know, going in that twenty twenty four, twenty five companies started to emerge out, but truly, you know, people thought AI back in the day was robotic process automation, not true AI. and twenty-five was a really interesting year because I think there was a lot of pressure from boards and executives saying, hey, we need to implement AI. We need to implement AI into our operations. And and they did that. Unfortunately, I I think you read a lot of the data today, says less than fifty percent of these AI solutions that they've incorporated have given the return on investment. And I think there's there's a lot of reasons why. And but at the number one, they just these companies were not AI readiness. There's a certain process you need to go to make sure your company's ready to absorb the AI. And it's it's not just techno technical. a technology issue and an API integration and knowledge management and all of those items, it's also a change management process. 'cause this definitely changes the way that you work internally within your operations, but it also is the impact on your customers. So

Yeah, just there's a lot of issues, but there's probably three or four major things that are very, very important to having a successful AI. One is there has to be buy-in across the organization. Right, from the frontline employees all the way up to the executives. And and that really comes around to change management because in in fairness to the frontline employees, there is a fear, right? That this is going to eliminate jobs. And for we have to be honest with each other, you know, tear one work, you know, the easy type of calls and stuff like that. We are having really good. impact of AI to be able to handle those calls. so I I think that that that change management's really important. Having, you know, having the right integration into your back end systems so the AI can actually do something. It is mind-boggling to me how many of these bots that are out there that are just if I was the CEO I'd just take it down today. I mean they're just they're horrible. I mean unless you want directions to the office or some basic QA, that they're not integrating these systems in. And therefore people get in and they try to use it for the first time. It it renders completely useless data and then even worse when they then pick up the phone and talk to somebody, they start all over again. That is just a really bad customer experience and stuff that we design around that we we just will not implement in AI. knowledge management is the foundation. You know, everybody wants to build the penthouse, right? And go to the top floor, but not n knowledge management is is what I call the foundation, the cement, right? So do you have a good knowledge management system? Because it's only as good as the data that it's pulling from. So a lot of companies need to really take a hard look internally at their knowledge management. Maybe they need to do a knowledge management transformation project before, or we can use some AI tools to extract all the knowledge across various depositories. And we go into organizations and they have knowledge in three, four, five locations or or they use SharePoint and And SharePoint, if you're in the contact center industry and you're an agent, I call it a search and rescue. I mean it's not a knowledge management. It's a great document repository, but it's not a knowledge management platform. So those are just three or four. There are obviously other areas that we have to look at from the cybersecurity and and and those type of points. But those are the three to four major areas that we make sure that the organizations are ready to go before we start a pilot because you're absolutely Right. I mean, the everything you're reading on whether it's LinkedIn or whatever, is that these POCs aren't working, they're failing. And we're sitting back here at Ascent Business Partners going, our last 64 projects, we have a 92% success rate. And what's really important is that how do you define success? Well We allow our customers to calculate the outcomes. So this is 92% of the time our customers are calculating that they're getting the return on investment. Our minimum, minimum threshold return on investment is 2.5x in year. That means for every dollar you invest on AI software. You will get at least two and a half dollars back. We're seeing much better ROIs, but that's the minimum standard because it's not very hard to go to your CFO or CEO and say, Hey, I've got this AI tool, I was able to prove it out, and this is a really important part with working with the Sent, is that before you make any commitment to the AI solution, You will have a POC with clearly defined metrics, and you will then at the end of the POC, which most of these are 30, 60-day POCs, they're not long POCs. you'll be able to prove out the data, but you'll have your people, your data, and your environment, and you will calculate out. So you'll have the raw data to go back to, you know, the executive committee and say, hey, we just didn't, you know, you know, buy this AI or had a presentation because how many of us have been on the back end of this where we went through an RFP process and we started with eight vendors and we got to six and four and two and one and we had beautiful PowerPoint presentations and great demos and everything. We buy the software, we in install it, and guess what? We don't get the return on investment. In our model, that'll never happen. You will have already put the software in your environment with your people, in your data, and you will have calculated the RRI. So it's a pretty, we call it a trust model. We trust our clients because they want to be successful. So we trust our clients to calculate the return on investment. And 92% of the time, and this isn't across five or ten projects, this is 60 plus projects, the clients have determined they're getting a minimum two and a half X in year, in year ROI.

Yeah, I I I think it's really important is that w everything we do is it is basically what I call responsible AI with employee in the loop. Really important. We're not going into clients and telling them, hey, we're just gonna rip out all your employees. That it doesn't work. We if I was starting up a new company, without a doubt, I would have AI as a integrated fabric into the company, right? Because I would be building a much more efficient organization. I'd have a huge advantage of starting up a new company because I wouldn't have any legacy systems that I'd have to work with. So if I could build a new company from the ground up with great employees and AI, I'd have a much more efficient, more effective, more nimble organization to move quickly. Because I mean at the end of the day, you know, if you really look where EIS had its impact. today over the last you know two to three years it's impacting the the BPO industry the outsourcing industry because companies typically would take their easier calls and they used to use labor arbitrage. I'll move them to the Philippines, I'll move them to India, I'll move them to South Africa, I'll move them to Latin because it was lower cost. Well those are the easy calls. Well those are the first calls that AI is impacting. So if you watch the BPO industry, the BPO industry in the last two years has lost Over $52 billion in market cap, right? Because you know, ATT would have 60,000 jobs that they would outsource. Well, in two years from now, they're gonna have 20,000 or 30,000 jobs that they're gonna outsource. So the biggest impact is being felt by the BPO industry today. and US-based companies, at least what we're seeing, and companies in AMEA, is that a lot of times they will reinvest. If if you were able to use AI to take out easy work, they would then reinvest that with these employees back into the company to do the work they couldn't afford to do, right? Do the real value added work for their customers, but they just couldn't afford it on the PL. Yes, they're gonna take some savings and some cost savings. So if we go into a client that has 500 seats in their call center, and let's just say conservatively we take out a hundred. They may let attrition bring down so no one's getting laid off. Through natural attrition, they'll come down, but then they'll repurpose maybe 50 of those employees. And that was very relevant on a recent project we did with a private equity firm where they had over 300 people doing scheduling. You know, it was a healthcare organization. And we, it was pretty easy work, so we had a much higher deflection rate with a conversational AI bot. with employees in the loop. But when they were done, they let some attrition, obviously for the cost savings, but they only reinvested half of the resources back into revenue generation. So not only now is the AI taken out cost and they're a much leaner organization, but their revenues are up 28% because they're doing value added for all the 2,600 offices that they're now supporting. So we're seeing a lot of that with the the AI.

Yeah. Well, I mean I mean that gets back to my early discussion that the stuff that was put out there was just it wasn't designed well and and again we're not gonna put our brand behind anything. Either we're gonna do it right or we're not gonna do it. And our customers wanna do it right. We just we've been doing this for four years and you know we're it's an interesting, you know, organization. You know, Ascent Disc Business Partners is a digital AI consulting company without a single consultant. 100% of our employees are operators that have owned PLs or technologists. So nobody grew up in the consulting world, right? Yes, we go in and consult, but we do it from an operational PL standpoint, you know, because we've been on the other end of this process. So we try to help our customers in there. But you know, the example you got is real. I mean, so if you had put a bot out on a website, 80% of it is just it's horrible. It's not, they haven't done the API integration. They haven't built in the use cases. There's a lot of things. And then even worse, what they typically do with bots on the website is then they'll put you to a live agent. But it's a chat agent, right? And what what happens most of the time with that chat agent? They have no idea what your experience was with the bot on the website, right? So we are we're you know we are moving for websites, fully integrated bots that that can literally integrate into an ERP system or integrate into basically their back-end system so that if you ask a question, you know, it it will be able to answer the question because that's ultimately, you know. Customers don't mind self-service as long as it's effective. But important in the design is that when they do get to the agent, the live agent will read the transcript and say, you know, hey Craig, I noticed that you were looking to try to do X, Y, Z. Boom, but how can I help you? Right? Makes the customer feel better. We do that same thing with the conversational AI. Because conversational AI has gotten so good today. That when you're calling into a center and you get a conversational bot, you don't really know that you're actually not talking to a human. Now, we talk about responsible AI, so therefore we want to ask our customers must disclose that they're talking to a bot. But we have certain conversational bots that that they'll take out and completely solve 60, 70% of some some call types. Other call types, it's 30. But when it transfers to a live agent, We have the customer on the phone, the conversational AI, and the live agent. And the conversational AI will talk to the live agent saying, I have Craig on the phone. He was calling us about this health benefit claim. Go through all that stuff. You know, I just I want to hand this off to you. So now the customer's listening to this handoff. They're happy. They don't have to repeat themselves. And by the way, we may not have deflected the call, but the amount of data that we may have absorbed, and if I have CD CTI integrations. The live agent has to screen up all the customer information, again, when designed properly, and then just picks up the conversation. But we've now taken 30 to 45 seconds up to a minute off that call. So although the conversational AI may not have completely deflected the call, they can take out a minute of productivity on that call and the agent doesn't mind that experience. The agent will go back the next time they call and they'll talk to that conversational bot because of that.

Yeah, it it's a good point. So yes, yes, we've we've used to charge consulting fees. And there are certain times if customers are asking us to come in and do an AI roadmap, not solve use cases. But but what I realized is that a lot of these companies budgets are really tight, right? And and going in and trying to sell a consulting project in an AI roadmap, you're one of a hundred companies that are approaching them. You know, it but you're when you go into talk to the executive suite and say, I am not gonna charge you anything. I am gonna be paid one hundred percent on the outcomes we deliver. And even more important, the trust model we talked about earlier, I'll let your team calculate our value. That makes it a whole different discussion because now The company still makes investment. It's not like free because they have to invest their people and their people's time is valuable, right? And that's part of the AI readiness, is making sure we have the right team and the right resources to make the AI implementation and POC successful. Because once you're in a a POC, remember our POCs are full production. So the i if the POC works and you have 500 people and you put 100 people into a POC, when you're done to roll out the next 400. Is days like to deploy it, the software is easy. Because it's already ever we've gone through everything you need to do to roll out a software in a full production environment. So this no consulting fee basically now takes out all the risk, right? I don't need budget now. I'm willing to put skin in the game, and 99% of consulting companies aren't willing to go in and put skin in the game. Now, I have a 92% success rate. If I was running a 30 or 40% success rate, I'd be bankrupt. Because the amount of effort we put into these POCs, these are real highly qualified 25-year vet resources, both operational and technology, that are involved working with our clients. If those clients don't get the 2.5X in-year, we don't get paid. Period. So again, I've been on the other side of this, you know, so we'd we're making it really easy. I mean, I've gone to my board and and and I've I've said 'cause they're running some of these guys are running really large companies and they're like Craig, you know, it's a high risk, high reward model. I mean, you're getting the results now, so you're you're it's a great model. It's a win-win, right? The clients aren't spending all this consulting dollars and they get to calculate the outcomes. Because the other part that we do is is is time to market. You know, we have invested millions of dollars of curating a network of partners. So if you tell me you've got a training problem, I'm gonna give you two or three companies. You want conversational AI for deflection. I'm gonna give you four or five companies. There are probably 250 companies out there doing that. You need to understand that. I mean, the cost of entry, there's every day, by the time we're off this podcast, there'll be there'll be 20 new AI companies that have been formed, right? Across the globe. I mean, everybody is it just this technology is so amazing that to build product. you know, from scratch is not that difficult. I mean we use AI within Ascent business partners. You know, we have AI agents. So I have I have I have over seven AI employees that do an incredible aw amount of work for us every day. And they're adding we're they're adding to our our current employees, giving us more and more capabilities as a company.

Yeah, no, I mean one of our our goals is right now we're primarily US based. We're now expanding into Europe. I I believe, you know, the European market has some more difficult use cases and challenges which we're we're ready to take on and but all of our solutions are GDPR, you know, they they they they pass all the InfoSec and all of the all of the technology stuff that IT has to go through before you're putting it in. But what's really next is is riding the wave. I mean, I don't think any of us have a crystal ball as to what's gonna happen with AI in the continued advan advancement. What we do know is that These products in this industry is changing on a daily and a weekly and a monthly basis, not quarterly, right? So one of our biggest challenges is having a dedicated team, just keeping up with the product development and also our research arm. We have a very heavy research arm looking for the latest and greatest. The our challenge is right now is that I think one out of every 14 companies that approaches a cent to get in the network, get in. Because I have CEOs that are coming to us saying, my God, I got this great product. And they demo, we go through due diligence, and we have to be brutally honest with 13 out of 14. I'm saying, you have a great product, but you don't you're not differentiated, right? There are 10 other companies out here, 20 other companies that are out here doing it. We we are looking now, because our network is so mature over the last four years. We're looking for product expansion, different areas, you know, outside, we're starting to breach a little bit outside. Of the contact center space and working on some areas around the contact center and back office automation and stuff like that. But at the end of the day, I don't have a crystal ball. I know one thing. It's going to change and it's going to continue to evolve, which kind of makes it exciting for us to be on the forefront of that. And then bringing those changes to not only new but existing customers that we work with is really exciting because at the end of the day, you know, if we can. continue to run a 92% success rate, which we're doing today. This is a big win for our clients. It's a big win for our CIOs and CTOs that will will their career wise. I mean, if they can point to use cases of successful AI when supposedly an industry is at thirty to fifty percent success rate and they're and they're running ninety two percent, it's a win win, it's a win from a cost, and I think the customers win. 'Cause if when AI is implemented properly, it is a much better customer experience.

Yeah, no, we we we we a lot of times we will talk to folks, we'll just have a twenty, thirty minute inter you know, call and w we will, you know, share with them, you know, kind of the model and we we our corporate deck, you're gonna love this, is two slides. So we have a we have a two-page corporate deck, and and most executives, the meeting starts off great when they find out we have two slides. That's it. Because it's a pretty simple message. Here are all the areas we can help you, and here's the process. We're not over overcomplicating this process. So so best way to reach me is my email. I I can drop it into the chat, or you guys, you could share that as part of the chat. I can drop it in, but or go to our website. But yeah, it the best thing to do is shoot me an email. We find a mutual time and we have a discussion and we'll see if we can help you. And like I said, there's no cost for a conversation. And if we move forward, we're not gonna charge you consulting fees. We're gonna solve your business problem.

Yeah, and I'm gonna I just put into the chat my email address. So feel free to reach out directly. Please mention the podcast. It would be great to know how you know how you got to us. And and we'd love to work with you. we're excited about our expansion into Europe right now. the US business is obviously continuing to grow. And we're just having a lot of fun, to be honest with you. This is this is really fascinating for someone who's been in this industry for almost 40 years now. the last three to four years has been so transformational that. if you're not living and breathing it every day, it's hard to explain to people, you know, the pace of change that's happening. But overall, if it's done well and you're doing it responsibly, right, and I think that's an important, you know, process, you can have a lot of success with AI. You don't have to be in that 30 to 50% that's failing.

Awesome. All right.