As artificial intelligence continues to dominate boardroom conversations, many organizations are discovering that simply adopting AI tools isn’t enough to achieve meaningful business results. In this week’s Weekly AI Update, David Lykken and Pavan Agarwal unpack why an overwhelming 88% of AI initiatives fail to deliver measurable value and explore what separates the successful 12% from the rest. Drawing on Pavan’s recent Forbes article and PwC research, this episode reveals why CEOs must lead AI transformation from the top, rethink entire business workflows, and focus on execution—not experimentation. If you’re looking to turn AI from a costly expense into a true competitive advantage, this conversation offers practical insights every business leader should hear.
[David] Listeners, we’re back with Pavan Agarwal with another AI update and are th ese ever being successful? I was just looking at the numbers, it’s really drawing a lot of attention and Pavan, I wanted to say thank you for doing these on a regular basis. A lot of people are tuning in, are learning about AI. So what is give us a little you’re you’re you’re on camera here. So those who are listening to this, he’s wearing a blue jersey. We were just back in Dallas. Pavan, talk about the blue jersey and why you’re wearing it.
[Pavan] Yeah, so we had a great day two weeks ago and we did the painted blue Dallas because Japan was playing for the World Cup and that day they were playing, I think it was Japan versus who was who was playing it that day? shoot. It was a tie it was a tie game. because I just came from Philly and I watched the France versus Paraguay and it was Japan versus I forget who it was, so anyways it was a tie game, so it was great day. But what’s more important is that the same day we had or the day prior, we had the Mindset Summit in Dallas at the Cowboys Club, which as you know during FIFA, FIFA takes over the whole Cowboys Frisco area. And we had to pull some strings to get the Cowboys Club given to us for the Mindset Summit. It was just a it was like a wow event.
[David] Is another wild event with speakers and every time I think you’ve outdone yourself on speakers or there there’s a new group that shows up and they’re even more dynamic. I can’t wait to have some lead on. and we you know there I want to talk about that, but you recently published an article last week in the Forbes magazine, and it’s titled The AI execution gap, why CEOs missed the mark and how to get it right. This is a really well written article, Pavan, and there’s some things I want to talk about. Why the title?
[Pavan] Well it’s because you know the this the article cites a PWC report that shows that only twelve percent of all of the initiatives that the industry is taking across the full spectrum, not just mortgages, but from manufacturing through tech, only twelve percent are actually succeeding in implementation of AI. Where by succeeding means you’re improving service and lowering cost. Okay, and for most companies they’re just increasing cost and in some cases worsening service. Okay.
[David] Yeah, by the way, I want to interject there because for some people don’t know what PWC we’re talking about there, it’s Price Waterhouse Coopers and it’s the report you’re referring to is the 29th Global CEO Survey. It was a big survey and very credible and what’s we’re you’re talking about, things are getting broken into two camps.
[Pavan] That’s exactly. And so the twelve percent is the CAMP A, which is actually succeeding with AI, and the remaining whatever, eighty-eight percent is struggling with it, and many of those will fail and there was another study, I think, by the Gartner report that says fifty percent of all AI implementations this year will be very expensive, right? very expensive relative to those companies and they will roll it back next year. So it’s a double whammy. One, you spent burn through a lot of money. Second, you have to roll it back. It’s sort of like I don’t know if you’ve seen those reports about Meta and I think it was Microsoft or somebody, that they had to put a stop on for restrictions on usage of AI because they have these huge bills of, you know, hundreds of millions of dollars, billions of dollars in a month because employees went crazy with AI and these massive data center bills they got, they didn’t expect it, right? And that didn’t translate to actual productivity. So they had to they had to put a put a stop on it. So if you don’t know what you’re doing with AI, you will burn through a lot of money. And so the bottom line of the Forbes article is that if you don’t know what you’re doing with AI, if it isn’t a well planned integral strategy and part of your company from the top down. That means that the C suite has to understand the limitations. Not the capabilities, but the limitations. Okay, and we’ve had a number of podcasts where I’ve talked about that limitations are more important than the capabilities because that way you know where to spend your money and where not to spend your money. Most importantly you know where not to spend your money. And if the C suite has to understand that and they and then they have to take a fresh look and retool their organizations to be AI based, right? Because you’re you cannot just take your existing workflows and slap AI on it and say that, okay, now we’re an AI company and now we’re gonna be much more productive and save all this money because all that happens is you end up creating more confusion and complications and AI is expensive and you end up burning through a lot of token cost.
[David] Yeah, well and what’s interesting in this article you said the data is sobering. It’s a staggering sixty-four percent of private equity and private invest principal investors in the C report says absolute they’ve had absolutely zero impact on the cost for their AI initiatives over the last twelve months. Eighty-two percent said no impact on revenue, which is what everyone’s trying to look for. They’re trying to get up overall only the twelve percent, the Vanguard group you were just talking about is successful achieving that. And you’re getting then you get into the trap of the AI or the experimental AI. What were you talking about there?
[Pavan] Okay, yeah. So you know, experimental AI to a lot of people, that’s like agentic AI and a lot of people listening to this thing will would have heard that term agentic AI that and then they believe and it’s marketed as if instead of hiring employees, you hire AI agents, in effect, quote unquote, hire AI agents. Which is large yeah, it’s for the most part it’s you know, it’s kind of true, but it doesn’t exactly work that way. It doesn’t it isn’t as if like I have somebody doing my payroll, and I’m going just replace that person with an AI agent, and poof, it’s that simple. It doesn’t work that way. It’s like with any technology, you know, there are actual limitations. And when you try to replace a person doing a payroll with an AI agent, you have to understand what the limitations are, and you have to redesign your complete, you have to re-engineer a complete process to be an AI-based process. And we saw a lot of that in the dot-com era where you have to re-engineer your company to be internet-based e-commerce delivery, right? That’s why Amazon is the biggest retailer in the planet and Walmart is number two. When you would have thought that something like Walmart should have should have taken over number one. And right? Because Amazon from the ground up is engineered to work for internet e-commerce. So it’s the same thing with you know, you take a legacy company and you and CEOs, C suites, they want a silver bullet. There’s no silver bullets, right? This stuff is hard, this transformative and it’s it is a re a complete redefinition of your organization. Okay, and to just to say, okay, go and tell the different you know, to kind of give a you know, top down directive, okay, and tell the different leaders within the organization, go use AI and see what you can do with it. That’s not a strategy. And that’s what everyone’s been doing. That’s not a strategy. And just to tell the leaders and tell the staff, go start using agents and let’s see let’s see how much money we can save. That’s not a strategy. Right? Yeah.
[David] Is this what you meant? You said you had a code. In my experience, many executives are stuck in what can be called pilot purgatory. By the way, I cracked up on that. They deploy AI at a service level bolt on. That is that what you’re talking about right there?
[Pavan]. Yeah. that’s right. And again, because there isn’t a directed com vision from the top, other than you go you know, we’re missing out, go take advantage of this stuff. And then so you have different team leads either buying pilots or buying pilot products from third parties or trying to set up their own pilots. Okay, and the big confusion with AI is you can create a pilot, right? What would normally have taken you know months and expensive engineering to create to get a prototype working, you can create that in an hour, maybe less than an hour. Okay. But now so when you create the prototype, you think, my god, I’m done. Well, what happens is then it’s then you spend the rest of your time trying to integrate it into your processes, trying to make it stable, trying to make it consistent and you end up spending a lot more time trying to take it from pilot to production. Okay. And people don’t understand that, you know, and I don’t blame them. Like, like if I if someone spend in you know 30 minutes, an hour generate a website and I’m like, hey, it’s done. Amazing. Okay. And you think that should be it. But there’s so much more to it, like, because you have to inner it that website that you created has to integrate with the enterprise databases and enterprise workflows and that’s where you end up in the purgatory because there’s these million other problems you gotta you gotta step through and then you gotta bring in new experts to complete that integration. So you know for example right we have something called a angel agents okay and which is basically AngelAI integrating with open claw and that gives you access to the massive global suite of agents. And we have a few a few of our partners, like all in lending, that is using it, that’s on our on our test environment, and they’re using it. And I just got a text from Chuck, I think it was this morning. He said, I built a web page in less than an hour, and he built it himself that he has never coded in his life, right? And he and he built his own website within an hour, right? and he’s able to do that and successfully move forward because Angel Agents is integrated into the Angel AI ecosystem, okay, and it’s talking and connected with your whole loan database, the whole loan process. So he can do that an hour and actually works and he can move forward. Okay. But if you’re an enterprise and you just tell your different people to go create, do stuff with AI and create stuff. There’s gonna be this massive problem when you when you take those whether it’s web pages or whether it’s a online ordering system or you know whatever invoice management or inventory management or whatever, and they had an AI agent do it. Now the problem is that okay, now I gotta integrate this with my database, and if this thing decides it wants to update my inventory balance or transfer money to a bank account, that’s really scary when real money when it touches it and tries to move real money around or real inventory around, which is real money, then then you that’s where the hard part comes to says, okay, how do we test this? How do we make sure it doesn’t ever mess that up? Right? And that’s the purgatory is that post prototype, right? Post production of the prototype, now you actually have to make it work with the enterprise. It takes you know twice as much engineering to figure out what they what the AI generated and how to integrate it smoothly and safely and securely. Right. And the other things you have to you have to really test. And there’s many reports that have come out about this where you know tech companies have been using agents to generate code and generate software and then they end up spending twice as much time with senior developers going back in and cleaning up all the security holes and all the things it it didn’t think about.
[David] It’s interesting. Yeah. Isn’t the whole objective to fundamentally alter the core workflow? You talk about this in the article, and it it’s supposed to reduce headcount in theory, accelerate core transactions, and reduce operational costs. Now you got that going on with angel AI, but we’re not seeing that happening with many of the systems. Only 12%. Is that what you said in the article or the study?
[Pavan] Yeah. The studies the PwC studies had twelve percent and I would wager that if we really dig into that, right, and we see where we stack up on there, we’d probably be in the point one percent of the twelve percent. We’re in the tip of that sphere. I’m not aware of anyone, any organization, any industry that has let their company run completely on AI. Right. We’re the only ones as and I’ve you know, you know I travel the world and I’ve searched high and wide for to see who’s doing what and trying to understand what what’s going on in the marketplace and no one’s figured out how to run an entire enterprise and delegate that entire process, manufacturing process to an AI. We’re the only ones that figured out and this is why we have so many patents on the product because we can do we can execute expensive high-stakes transactions with the AI and that’s what you need to run it run a production facility, right? Like for example, if you were to use an AI, for example, to control your oil refinery, okay, then it needs to understand that if I turn this nob the AI says I turn this knob and I increase the amount of gas going into this pipe, I could blow up the plant, right? It needs to know, you know, when to increase or when to decrease. Are you gonna let an AI make that control automatically? Right? And if it gets if it if it guesses wrong, if it gets if it gets it wrong, your plant, you know, is gonna blow up. Right? And and the same thing in the mortgage business, right? If you’re letting an AI make an underwriting decision, if you’re letting an AI say, look, we can exclude this debt or or you know, we don’t need another bank statement or only need one year’s tax return, I don’t need two years’ tax returns, whatever, whatever it is, if it gets that wrong, guess what? You gotta you’ve got a repurchase on your hands. You you’ve got to spratch a net loan. You’ve got or you even worse, you could end up denying a borrower that deserves that credit. Okay. Right? So this is what a deterministic AI means. And what it means is when you design an organization with the clear vision of an AI based production facility with an AI management. So you gotta have the vision and then you have to have the technology and you have to have people who understand the how to use the technology, understand limits of technology, and then your entire workflow is built with that technology.
[David] Is this what you’re talking about when that part of the article you say building a transactional revolution?
[Pavan] Yes, exactly. A transaction revolution is everything I just explained, which is an AI that can process transactions. That just like your self-driving car, you know, you don’t think about your self-driving car as a series of transactions, but it really is. Right. In computer geek terms, it’s transactions because you have you have some input, something in the camera. Camera senses the dog right in front of the car. Right. There’s a transaction that happens at that point. The data is there’s an obstacle potential obstacle in the car and the decision is to press the brakes, or the decision might be to veer right or veer left, right? That’s a that’s a single transaction. Right. And you gotta be able to make these real time transactions you know in in microseconds with data coming at you, you know, like through a fire hose, like so much data coming at you at the same time and you gotta be able to make these decisions really fast. That’s the transactional revolution right there.
[David] Yeah, and well, I mean, I’m thinking about the Jennifer interview that I it says the pattern that changes AI forever from probabilistic to warranted intelligence. Now, your article you talk about generative AI to deterministic AI or decision intelligence, but it is actually in the case of angel AI warranted.
[Pavan] Correct. Correct. So warranted is that’s essential because when you as a mortgage broker or as a consumer, as a realtor, go to Angel AI and you give it pay stubs and you say and you say you upload some pay stubs and say, Hey, based on these pay stubs, what is my qualifying income? And if Angel AI comes back and says your qualifying income based on these pay stubs is you know five thousand dollars a month, okay, that’s we’re gonna stand behind that 5,000. It’s not like there’s not gonna be anybody, a QC department or an underwriter. There there’s no more humans are gonna look at that. Okay. And we’re not gonna say, oopsie, no, we meant 4,500. Sorry, gonna take that back. And everyone who’s watching this, everyone, every mortgage broker, everyone’s originated a loan, was watching this. I guarantee you it’s happened to you more than once, many, many times, where you you’ve been told based on income documents or bank statements or whatever that we can do X, that the income is five thousand dollars a month, and then when it goes to underwriting, right, and the underwriter reviews it and she knocks it down to 4,500, and now your whole deal is blown up. Right? That has happened so many times. It happens every single day to loan officers all over the country. But in this case, you give the AI the information and in minutes is going to come back and say, here’s a qualifying income and then that’s it. You don’t have to second guess that because there’s an ask there’s real money behind that, there’s a warranty behind that, and I’m gonna honor it. Right and that changes the whole world and the whole way you do mortgages. Cause when you actually have dependable, reliable, you know, when you have a when you have a process where you can count on the words that are given to you, right? When you have an AI that you can bank on, right, now you have to rethink and retool how you do this business. Okay. So one of the one of the conversations we have with our clients all the time is we’re saying like you’re an AI originator now, right? The how you do things has completely changed. Just go out and get more business. You know, ’cause the classic workflow is you interview a borrower, you take you take documents, then you hand it over to your processor, and then your processor reviews it carefully and thinks about it and maybe asks the borrower for more information and prepares as perfect a file as possible before submitting it to underwriting because the processor doesn’t want to get penalized for having conditions from underwriting. Right? And that’s like totally, you know, asked backward. You know, excuse my language here. Right? It’s like, why are we asking processors to be underwriters? That’s you know, if asks processors to be processors. What that means is go and communicate with the consumer, right? Help them understand the technology, communicate with the consumer, keep them happy, keep them excited, keep them motivated, and keep them sending the documents that are needed and as soon as you get a document, put it in the AI, let the AI in Minnesota tell you what you know what the income is or what the balance you can use or what the ratios are, or whatever it is, right? Let the AI tell you this, that whether it’s acceptable, not acceptable and you just keep on going. So let’s go back to redoing workflows with AI. This is kind of you know, I’m talking about this, I’m talking about what processors are doing wrong, what brokers are doing wrong, being AI brokers, because it’s it and that speaks directly to the heart of what I wrote in that Forbes article is when you’re an AI company, right, then all the old ways of doing things go out the window because and the reason is in the old ways of doing things without AI, you have all of these restrictions on you, all these training wheels on you and now you’ve been doing it for so long you gotten used to being confined in that box. Okay. And when you when you’re doing AI, all that that box is gone and you’re free and you can fly, right? And you can fly out of your cage, but you’re still staying in your cage because you think the cage is still there, right? Even though it’s been it’s been removed. Okay. And so you know because like we’ve had situations where brokers have put in loan documents or you know credit documents into the AI and AI came back right away and approved it or declined it or what you know, whatever, made decisions and told what to do, right? And they’re like, it’s broken. Well like why do you think it’s broken? ‘Cause it came back too fast. There’s no way it could have done this this quickly. Like, no, it’s done. This is what you need. You know, do this and let’s close this loan file. Right. So right. So yeah. Yeah, exactly. So big adjustment and you gotta redo the way you have to redo the way you change sorry, you have to redo your entire process workflow, right? And you have to reprioritize what’s important for your staff, right? Right. So for if you’re if you’re in a mortgage originator, what’s important for your loan processor? It’s no longer keeping an underwriter happy, right? Because now an un now the underwriter is AI and it’s not going to be unhappy or happy with you. It’s just it just wants the information and it’s going to tell you right away. Right? And you don’t have to worry about keeping the underwriter happy and because underwriters get grumpy if you piecemeal it to them. Right? Well this is an AI. It keeps track of whatever. So as soon as you get something given to the AI, the AI is going to review it on the spot because it it always has the information about all of the loan in his mind all the time. So you get something, give it to the AI. It’s that easy. So that you can focus on communicating with the borrower and making sure the borrower is motivated to the whole transaction, right? And that the borrower has a good relation. The processor’s main job is making sure that the borrower’s satisfied and has a positive opinion of his loan officer and a positive opinion of his realtor. That’s the number one job of the processor. It is not to make sure the processor’s job is no longer to make sure that the underwriter has a good opinion of the processor, because that’s the old paradigm. Right? The old paradigm processors are completely focused on making sure the underwriters are happy with them so that they don’t get in a penalty box. And the new paradigm, the processor should be completely focused on the customer satisfaction of the borrower and the realtor.
[David] Totally makes sense. You talk in this article about too many CEOs are delegating their AI strategy to their IT departments rather than leading from the C suite. And they’re treating software. I mean, talk a little bit about what you’re seeing so many of these people. One of the things I audience I won’t I really enjoy about Pavan as a human being, he’s not just talking about the success he’s having, Angel AI. He’s generally trying to help the industry roll out, implement AI in an effective, intelligent manner. So it’s a big mistake. I’ve seen it happen where they say, IT, get this. There you talked about business process, how we have to approach this from re engineering our entire workflow the way we go about the business.
[Pavan] Yeah, I mean, let me think of a good example. If you if you’re a shipping company, okay, you don’t just tell your procurement department go buy some trucks, right? because they have no know, ’cause you can have your head of procurement, your chief procurement officer, giving the instruction to go buy some trucks. Okay, he’s not gonna execute that well. Because that requires vision from the chief executive officer because he’s running the he understands the big picture of the shipping company. He knows what kind of trucks he needs. He knows how many he needs, right? He knows he knows how far they’re going and all of these things, right? He knows the limitations of his business, right? So you work with your chief procurement officer and you work collectively together and you make if you have to start with a vision of this is what I’m gonna do with the trucks. And then you work with the chief procurement officer so that you can get the right trucks to run your business effectively. Okay, the same thing. The C-suite can’t just go to their CTO. The CEO cannot just go to the CTO and say, Go give me some AI. Because why? Because everyone else is doing it. That doesn’t work because the CTO, as brilliant as he may be, He doesn’t have the same bird’s eye view that the CEO does. He doesn’t know the vision, the direction. He doesn’t know what kind of AI. What do you want to do with it? What are the goals? What workflow are you trying to affect? And what you know what metrics do you are you want to deliver? without that kind of understanding, what is your what is your immediate vision and what is your long term vision? Okay. And it isn’t as simple as, let’s just go buy some AI and throw it in there, right? That it doesn’t work that way.
[David] Yeah, it’s so important that you have a clear vision on this. And most people are not getting down to a complete re-engineering of the whole process.
[Pavan] Exactly. Bottom line, there’s no shortcut. And if you’re a CEO listening to this, you gotta go do your homework. You gotta go learn this stuff and understand it yourself so that you can you and when you understand the fundamentals and understand how this stuff works, you can construct a vision and you can say, Hey, with this with this, I’m going to this is the vision of my company, how I’m gonna do it. And then your CTO comes in and cons and builds it out to your vision. Right. So there’s been a number of articles, a n number of, you know, one man, two man billion dollar companies with AI, because the one or two people that started those companies started with a clear vision that I’m gonna use AI to deliver this kind of service and they realized that they could just build the business you know, with AI agents and start started doing volume. The name of the company, there was a company I just read about it. The name escapes me, but this guy, one man show, and he had a billion dollars in sales in in a quarter. And because of he had a clear vision. He had a clear vision and he understood the limitations of AI and understood what it could do and implemented it and delivered it to his customer base.
[David] Wow. Amazing. In the article you talk about shifting from vanity metrics to value metrics.
[Pavan] Correct. Yes. So vanity metrics is again, you know, having bragging rights that I’m an AI company, I’m AI forward, right? What whatever, I don’t know what the latest buzzwords are. And vanity metrics is just like, I’m an AI company because I just spend a billion dollars building data centers. Right? I’ve seen statements like that thrown around. I’m an AI company because we burn, you know, a million tokens a month. Those are great vanity numbers. those are great you know, for the uninformed, it sounds cool. Right. But actual numbers like real world metrics means and business is business, that means if I spend a million dollars on AI today, did I get at least a million dollars of value? Right? Do I get a million dollars more in sales or did I cut a million dollars in cost or a combination of the two? If those are real numbers and those are harder to achieve and as this report shows, eighty eight percent of the companies are not achieving those numbers, right? But they’ve spent all this money, so they’ll focus on the vanity numbers as opposed to the hard metrics.
[David] Yeah, that’s really good. of the things I enjoy about you is your care for people. You care about people, whether it be the consumer, also the worker. You said in this article, do not simply aim to eliminate a percentage of your workforce to save money. And you say, in your experience, Pavan’s experience, true power of AI is in how it can free up humans to do what only humans can do, build relationships.
[Pavan] Exactly.
[David] Navigate complex. I mean, I love that that principle by which you operate. And you do this with your own business.
[Pavan] Exactly. Like for example, I mean, our underwriters, right? I mean it’s been a multi-year decade plus long process. They’ve all been re-educated to become AI data engineers. So they spend their, you know, when their workflow is and they and when they work on exception, when they work on a document that’s sent to them, like, hey, you know, AI couldn’t make a decision on this, right? And they review it, they analyze it. Okay, and then they submit it back. Okay, what they’re submitting back is a data analysis, right? They follow a process so that we get we get the data back and so the AI can learn from that decision process. Right. So now they’re actually training the AI as part of the normal workforce. And so we didn’t have to lay off a single underwriter as a result, because now their knowledge is going into making smarter and smarter AI.
[David] That’s amazing, in the article, Pavan, you talk about AI should be a massive profit center, not just an IT line item cost. Rather than seeing AI can write, marketing, copy, consider utilization, utilizing it to identify, engage, and structure deals for a proportion of your new net leads.
[Pavan] Yeah, yeah. Correct. Correct. So you know, number one use of AI in the real estate and mortgage world right now is the people are using it to create marketing, whether it’s flyers or social media posts, or even you know, copy for websites and things like that, which is great. I mean it’s a great way of it saves a lot of time. And the and if the English and grammar is always right, which is beautiful. although you know I caution you because I’ve seen enough of it where I can see AI slots versus actual creative work, right? ‘cause AI generated copy starts becoming sloped after a while. so be careful with that. but nevertheless ne nevertheless, if you know those things are additive, they’re assistive, okay, but you gotta take, you know, and those things are kind of like in a copilot function they can help your marketing teams out, you help sales teams out, they get a little bit this takes some of that that tough grunt work out of their hands, which is great, which makes your sales team more productive. Okay, but that’s
[David] Well brings better yeah, greater job satisfaction and because they’re really like they’re they are being used more than just cogs and a wheel. They start actually com contributing the more important parts of the transaction.
[Pavan] Yeah, yeah. Correct, right? So it’s a good thing and keep doing that, but you you’re gonna see that the effect of that is on the margin. And you’re not gonna get big lift in, you know, are is that really increasing your sales numbers? Is that really i cutting your cost down? And the answer is gonna be no unless you re engineered their work process to be AI centric. That means have like like for example, our sales team, right? Our county’s ex sales team, they’re trained now by AI. So now we have AI that’s training them on a regular basis on what to say, what not to say, on you know complex scenarios that gives them the answers and so forth, right? So they’re becoming better professionals because the AI has the knowledge to direct them correctly. Okay. So you see the shift where instead of the AI being a co-pilot where you get an instruction from the salesperson and it goes and generates an email or something, now the AI becomes the pilot and is giving the salesperson instructions that hey, treat this account this way or you know, communicate like this with this kind of a problem, right? You see how the paradigm shift has to happen and that requires a culture shift. That means you have to train your salespeople. Again, I’m using sales as an example. You have to train your salespeople to accept direction from an AI, which is which is you know, a lot of organizations are gonna struggle with that. Okay, so that’s a very simple example of being AI first, right? Being AI centric. but a more broader example is like actually manufacturing, actually, you know, letting the AI process and underwrite the loan, right? And again, going back to the same example, you know, teaching your processors that you gain nothing, and in fact, you only lose time and you cause delays and create stress if you try to create that perfect file for a underwriter, a hypothetical underwriter who doesn’t exist anymore, to tell you to give you a attaboy, good job, that was a clean submission. Right. There’s the whole concept of a clean submission is obsolete now. Right? Just get the paperwork and let the AI do its thing. And as you get stuff, let the AI process it and it builds the file for you. It is it is your processor underwriter at the same time. You so it’s a complete change in in in psychology and how you think about the workflow. I mean this is an entirely different workflow.
[David] Yeah. I mean it comes back to workflow and rethinking it and understanding it. But yeah, listeners, I really encourage you to read this article. I put a link to the article that Pavan wrote in Forbes magazine, respected business journal, that talks about this. Pavan, great article. Always fun to have you back on the podcast. Kudos to your success. I love how you put people first and putting this into a proper perspective where AI is driving profits and making a difference to free people up to do what’s most important. And that’s better serve the customer. Pavan, thank you and kudos to brother.
[Pavan] Thanks, David. Cheers.
Important Links
Pavan Agarwal is a renowned leader in the mortgage lending industry and a pioneer in bringing artificial intelligence to the financial markets. Agarwal serves as the President and CEO of Sun West Mortgage Company and Celligence International, LLC.
