Recently, I had the pleasure of joining Dr. Christopher Loo on his podcast to discuss a topic that’s very close to my heart: transforming the mortgage lending process with AI. It was a fantastic opportunity to share insights about how our AI loan origination software is revolutionizing the way banks and non-bank lenders operate. Here’s a recap of our conversation, highlighting key takeaways for anyone interested in fintech, AI, and the future of mortgage lending.

What is Addy AI

Addy AI is an AI-powered loan origination software that helps banks and non-bank lenders automate the loan process. Our tools assist loan officers with due diligence, application reviews, and moving loans from application to closing. Our software is already in use by major lenders, processing numerous mortgage loans efficiently.

Addressing the Pain Points in Loan Origination

One of the core topics we explored was how AI is streamlining the traditionally cumbersome mortgage lending process. The conventional loan origination process is often bogged down by manual due diligence, extensive documentation reviews, and slow decision-making. By automating policy checks and correspondence, AI significantly reduces the time and effort required by loan officers, allowing them to focus on more critical aspects of their job. This not only speeds up the process but also enhances the accuracy and fairness of loan approvals.

AI’s Role in Making Mortgage Lending Fair and Efficient

AI’s role in making the mortgage lending process more equitable was another crucial point of discussion. Traditional credit scoring systems and manual reviews can often be biased or inconsistent. AI, when properly trained with diverse data sets, helps eliminate these biases by ensuring decisions are based on comprehensive and equitable data analysis. This is particularly beneficial for non-traditional applicants, such as entrepreneurs and independent contractors, who might otherwise face challenges in the loan approval process.

Enhancing the Client and Lender Experience

From the client's perspective, AI brings a host of benefits. Faster response times, personalized policy explanations, and interactive AI assistants make the loan application process less stressful and more transparent. Borrowers can now receive instant feedback on their applications and understand their mortgage terms in simple, natural language, rather than wading through pages of complex legal jargon. For lenders, AI addresses the issue of being understaffed by equipping loan officers with tools to handle more applications efficiently. This boost in productivity helps meet the increasing demand for mortgage loans without compromising on thoroughness or accuracy.

The Future of AI in Mortgage Lending

We also explored the future potential of AI in the mortgage industry. While AI can’t yet replace the nuanced judgment of human loan officers, it acts as a valuable co-pilot, enhancing their capabilities. As AI technology advances, it will continue to take on more complex tasks, potentially transforming the role of human workers in this field.

Stay Connected

The podcast was a great platform to discuss how Addy AI is transforming mortgage lending with AI. Our mission is to make the loan origination process faster, fairer, and more efficient for everyone involved. If you’re interested in learning more about Addy AI, please visit our website Addy AI.

Thank you for reading, and I look forward to continuing the conversation about the exciting future of AI in mortgage lending!

Watch my conversation with Dr. Christopher Loo about how AI is changing the mortgage lending industry

Full transcript of my chat with Dr. Christopher Woo

Dr. Christopher Loo: Hey, guys, welcome to this week's podcast episode. And I've got a really fantastic guest for you today, Michael Vandi. And today's conversation is all about AI and loan origination, which is going to be really interesting in the fintech space. He is a CEO of Addy AI based out of San Francisco or in the same city.

And I'm really happy to welcome Michael to the show. It's gonna be a great conversation. Welcome, Michael. Thank you so much. Happy to be here. Yeah. I'm always scouring the globe for, you know, just the innovators and disruptors. And so you're one of those. And so talk about your journey. , what , Addy, AI Addy is, how the audience can learn more about you.

Thank you.

Michael Vandi: , so I am the CEO of Addy AI. We are what's known as an AI loan origination software. So what we do is we help banks and non bank lenders train AI assistants to automate the loan origination process. So helping loan officers do things like due diligence with AI, automatically reviewing , applications with AI and then moving it from phase all the way from like application to closing the loans and then our software is being deployed at one of the largest, , you know, non bank lenders in the U.

S. and then they process, , a lot of mortgage loans, , a year.

Dr. Christopher Loo: Yeah, and it's quite interesting because, , I've been doing a lot of, , you know, I do a lot of research and thinking and I just this whole idea where you, you know, this whole like loan origination process for housing and, you know, these credit scores are really, , they're really, .

Nonsense. And, , you know, this whole process is just so c bersome. And it's just, I mean, only like, only if you are rich, or if there's like, you can have access to this. So talk about, you know, this idea of where Addy AI came about. And, , you know, we'll delve into more of the specifics.

Michael Vandi: Yeah, so there are a couple of things.

Sometimes it could be less c bersome if you look at like what kind of data they're looking at. And sometimes it could be very c bersome. So if you're looking at, let's say, bare credit scores, if your credit score is not over 50, it's like an automatic rejection. So it's easy to write software for that.

If you need to have like a certain amount of assets in your bank account. So when you submit your bank statement, it's like an automatic rejection. , now where the c bersome process comes into play is when they have to do due diligence into the entity's background, whether it's like a company that needs the property, or it's like an individual that needs the property, the loan officers have to spend days and weeks, and it's like a lot of back and forth, and anyone who's gone through, , the mortgage lending process, it's like a A really long and stressful process.

So giving the loan officers like the tools to make the process faster, , is what we, we try to do at, at Addy AI.

Dr. Christopher Loo: It's quite interesting. Cause I hear these stories, like for example, the, , was it, , , it was, , the Apple CEO and, , he, you know, he, he was able to apply for, you know, credit very easily.

And then his wife. She applied for the same thing and she got rejected And you know things like I like one of like I read on twitter There's one guy he paid down his balances and his credit score went down and he's like wtf, you know, it's like You know, it's like it's supposed to go up but you know Just all these Crazy, ridiculous stories, but, , you know, , so talk about, , one thing is, , talking about, , how AI is changing the mortgage lending industry.

How is it getting rid of these inefficiencies? How is it making it more fair and equitable? , I'm really, I'm curious.

Michael Vandi: Yeah. , so when you look at like, , any, any sort of industry, right, In order for it to be efficient at any stage of the process, there has to be a clear path to the next stage. And, , think of things like production lines at car companies, you know, after the doors have been fixed, it then goes to the electricals and then things like that.

Now it's a little bit different in the mortgage industry because sometimes every single case is different and so people have to find the time to like figure out, oh hey, but what's the policy in this particular state? Like is there some sort of like state policy that we're meeting and things like that.

So one of the first things that , AI is changing is doing all these automated policy checks. So as these scenarios come in, it says, Oh, first of all, the location is in Maine, we don't lend in Maine, so that's like an automatic disqualification. So a loan officer doesn't have to spend like 20 minutes going through the doc entation to see that, oh, hey, we don't even do this.

And then after that, it's like all the little steps in the process wall. What are some of the information that we need from this, , , borrower or like client or, , or broker? And then you would want, , to have some automation in terms of like helping the loan officer do that back and forth communication until all the information is present.

And then when the information is present, then the loan officer can actually be like a h an reviewer. , that's how we mitigate risks. So we don't automatically say, Oh, hey, once you have all the information, like, hey, you're good to go, we try to think of it as like a co pilot that you give to loan officers to help them make decisions better and faster.

And, , obviously, it's like a great experience for both the clients and the vendors as well.

Dr. Christopher Loo: Yeah, so so so I love this. It's basically using data and algorithms just kind of cut out all the excess BS. And then, , so the other question is, , and I love this idea of you have like a h an at the final stage, kind of saying, Okay, all this has been now.

Now we need to just kind of do a check, you know, things that really need to be checked. So, , we'll talk about, , the other question I have is, , so, , you know, Basically, , we talked about is changing the, , , mortgage lending industry. And then how are you looking at, , the, how AI is changing the way home loans are reviewed?

Like, what is like, you know, the typical processes, you know, you gotta, you basically get pre approved and you submit all your, , forms and all of your income statements and it gets reviewed. , how is AI changing that? Thanks.

Michael Vandi: Yes. Well, one of the ways that AI is changing it is the process is becoming a lot faster, so I can speak of it from the perspective of the, , broker or the clients, and then, then I'll speak on the perspective of the, of the lender.

And so, in the perspective of the client, , you get to receive correspondences really, really fast, then you don't have to wait for a, , loan officer to do their research before they get back to you. So you get to have like a great experience. Another thing that's changing is you get to understand your policies in a personalized manner.

So when you get approved, and then you finally receive that policy, it's like dozens and dozens of pages, and you don't know the fine prints and things like that. And then you can actually chat with a personalized AI assistant and say, well, What is, what am I getting into? And you can chat with it in natural language and it tells you exactly that, Oh, Hey, this is like a 30 year mortgage.

, you have to do X, Y, Z. And then if you don't pay X, Y, Z by this time. So it explains to you like a h an, it's not like, The huge jargons that you don't have to understand. And so that leads to like a great experience for, , the broker and the client as well. And there's so many different kinds of, , mortgage loans, like, and then the ones that you could qualify for, let's say like VA construction loans, which one of them have different policies that not everyone can understand in their head.

Right. So having this assistant that can tell you about it is really great. , and then on the side of the lender. , we all know people are understaffed. And so what happens is these loan officers are overwhelmed with the n ber of applications that they have to review. And so these lenders have the possibility to equip the loan officers with the tools that they need to scale it much faster.

And the bottleneck in most of these systems, the bottleneck is, , people working slower and productivity gains and productivity increases. And that's because there are more people, more entities applying for mortgage loans that there are loan officers to review these, these loans. So it's like just supercharging them with all the tools that they need to make decisions faster.

, and one thing I would like to say is that the way I see AI changing in financial services is that because financial services is like very n ber driven, , , it has a qualit, , quantitative measure. Yeah. And so with chat GPT when it hallucinates, , you don't know whether it's right or wrong because you need to have like some expert fact check, right?

, with financial services, there's actual n ber. There's a source of truth, which is a n ber that you can actually check, , if, if it's right or wrong. So that's one of the ways I think AI is changing financial services and mortgages.

Dr. Christopher Loo: Yeah, the other question I have kind of follow up question is this idea because that was a lot of my friends and colleagues, , you know, we're entrepreneurs and, you know, we've transitioned from this idea of W two to more just, , like building businesses and we were just.

We're out of the, you know, shooting a shit with over beers. And we're just talking as like, how does the algorithm treat, you know, entrepreneurs and people creating businesses over w two? Cause in the past, you know, you had the, some of your w twos. And now, now if you're like, , consultant or, , you know, , independent contractor, you know, you submit 10 99s, how, how is this?

, and it seems like, you know, for business owners, , that make more than w two, the, , system. This favors them over W2. So how is AI changing that?

Michael Vandi: Yeah, so the good thing about AI is you can train it. And, , well, that's also the bad thing about AI, because when you train it with, , flawed data, then you get, you know, garbage in, garbage out, right?

And so we try to train with a diverse amount of, of, of data, and so there's this thing called overfitting in AI. Which is when the model becomes trained so much with like one particular data set that it becomes skewed to that data set. So if you train it with a bunch of, , data from like, , let's say contractors who are not on W 2, they have companies and then they get approved a lot.

And the, the tool learns that, hey, these are actually people that we should be lending to. But if we try to not skew the data, , then it becomes fair and equitable. And so we have done a whole lot of work in trying to make sure that our training data is not skewed towards a particular group of people.

Dr. Christopher Loo: Yeah, so basically your training. Another Question I have for you is because you mentioned this idea of more applications to loan officers. And so what is this? You know, basically, we've seen it with, we've seen it with Uber and Lyft, basically, and we've seen it with Netflix, just basically, and now we're seeing it with also Tesla, just the obliteration of just The need for h ans.

So what is this going to do to the h an side of the equation in the mortgage origination process? Wow, that's, that's a tough

Michael Vandi: question. And my honest opinion of like, where did technology is right now?

Dr. Christopher Loo: Yeah,

Michael Vandi: it's where it is right now. It cannot replace every single thing that a h an does, and you actually would not want it to at this moment.

, you would not want to give, so the majority of Americans have their net worth in their homes, and you would not want to give that power to a machine deciding the asset class of the majority of Americans, right? And, , how we try to think of these things is, , as a co pilot for, , these loan officers.

So, one thing that we've seen repeatedly is, well, when certain models come out, they are as smart as, let's say, a 6th grader. And then After a couple of months, they are as smart as someone who's in 12th grade. And now we're getting to a point where they are as smart as someone who's, , you know, as a new grad intern coming in to help you.

And so if you try to think of it as a new intern. Who you can just give particular tasks to we have internet work and we give them all the Scott work to do. Imagine being like a residency, you know, so we try to consider it as that. So like the loan officer right now is like the lead surgeon and then the AI is a resident right now.

And over time, , 10, 15 years from now, that will change. And when that changes, we have to think about, you know, how we approach it as a society. And that's not a question for someone like me, who's an innovator to think about. , well, I should think about it, but that's not a question that I should, I've been making policies to like effect.

Yeah,

Dr. Christopher Loo: yeah, I love that. I actually think, you know, I actually think the obliteration of industries is actually good because it forces people to get off their butts and, you know, kind of do things, you know, as opposed to, you know, because like, I go around and I see all these like, government workers, and, you know, they just say, you know, Sit there and basically do nothing.

And it's like, you know, you could so it's good to have like this disruption, you know, for better or for worse. But, , and, , the other question I have is so, you know, you're in San Francisco, one of the best places for VC and based. You know, startups, you know, kind of the home of all the, the mag seven, , talk about just kind of, , you know, , your journey, , why you started it, you know, kind of the pitfalls and challenges and then end it with, , how people can find you and, , follow you, et cetera.

Michael Vandi: , I have an unconventional journey to Silicon Valley and tech. I was born and raised in Sierra Leone, Freetown, Sierra Leone, and from a very early age, I've always loved building things. I learned how to code when I was, , 14. My mom was a teacher. So she introduced me to all these technologies. , I remember when I was seven years old, I think I got my first laptop and I might have been the only kid in the neighborhood might have been the only CD with a laptop at seven years old in West Africa.

I learned how to code got a desire to build things. And then I applied for college in the United States. And then I, you know, came to the US. So I studied computer science and then I was involved in like a bunch of extracurriculars, went on to work at AWS on AI, and then I did my master's in software engineering and AI at Carnegie Mellon.

And then, , while I was in school, actually, I started tinkering with, with AI. And, , I've lived through two different realities where in Sierra Leone does not a whole lot of mortgage tech. In fact, there's like no mortgage tech at all. And then I come to the U S and it's, the difference is like night and day.

And when I started AI, I think, well, this could be improved even more. Like the processes are so mundane and slow, they have a bunch of roadblocks that could be improved even more. So I started tinkering with, , creating. , this tool that could help, , loan officers, , with email correspondences, and then I kid you not a bank reached out and it was like, Hey, this could be really useful.

And I didn't even intend it to be a company at the time. I was like, well, let's chat. And then we started chatting. I built a small team of my friends from, , Carnegie Mellon and, , it became extremely useful to them. And once I graduated, I was going to go back to Seattle to go to AWS. And I was like, well, I don't need to.

I could just. Stay in San Francisco and, , you know, move to San Francisco and work on this company. And that's what I did. And we've been able to do really well ever since. So it never became as like, I wanted to build a company. I was just like tinkering with what could be improved. And then a customer that needed it reached out and I learned about the industry.

Dr. Christopher Loo: Yeah. One, , one question I have just personally is because I see a lot of innovation in London and Dubai, Singapore, , you know, of course, you know, China, , and eventually India. Do you think do you think San Francisco is still the best place to, , for tech startups? Or is it, is it kind of shifting away?

Now we have, , you know, we have zoom and remote and, you know, a lot of the, , investments are going outside of the United States. I'm just curious about your thoughts.

Michael Vandi: 100

Dr. Christopher Loo: percent

Michael Vandi: Cisco is the best place. And I don't say that to build a tech company, not only a tech company, but a software company. I don't say that because I will obviously I am biased, right.

But as an entrepreneur, the odds are stacked so much against you. That the, the, the, the most minimal optimization that you can do is like where you live and the location. And in San Francisco, there's like a lot of resources and networking opportunities around. So 99 percent of businesses fail. And then the list that you could do is move to a place where you have more chances of success.

Now, Dubai and other places do have chances of success. Absolutely. But in San Francisco, it's like unparalleled and, you know, it's coming back. I'm glad you're in San Francisco. So, , if you want to chat sometime, go hang

Dr. Christopher Loo: out. Yeah, yeah, definitely. Like I said, , you know, I, I'm in San Francisco for the s mer just to have more, , exposed to more stable climate.

, you know, and, , so, but, , yeah, it was really fascinating. And how can people, , contact you, find out more about, , , Addy AI, , all your socials and check out the work that you're doing.

Michael Vandi: Thank you. Our, our website is addy. so that's A D D Y dot S O. , they can go on there and then contact me. I'm also on LinkedIn, Michael Vandi.

Reach out to me on LinkedIn. I'm very active on there, and I'll be happy to chat.

Dr. Christopher Loo: Yeah, and I really enjoyed this conversation, and I always love talking with tech entrepreneurs. I'm very forward thinking and just discussing these issues, and for all the audience, be sure to give Michael's socials a like and follow it.

Check out Addy AI, see the work, and thanks so much for coming on. Thank you so much.