372: What You Need to Know About AI

Buck: Welcome back to the show, everyone, today. My guest on Wealth Formula podcast is Shashank Shekhar, founder of Insta Mortgage. After starting his business in possibly the worst year for financial markets. 28, Shashank has led the company to be one of the fastest growing mortgage companies in America, helping thousands of families secure better financing for their homes in 2017 and 2021. InstaMortgage, which is formerly known as Arcus Lending, was named to the INC 500 list of the fastest growing private companies in America. Shashank, welcome to the program.

Shashank: Thank you for having me, Brooke.

Buck: So, you know, first of all, tell us a little bit about your company in that. And, you know, like I think one of the things I really want to get into is the technology aspect. But tell us a little bit about your company and what makes it different.

Shashank: Yeah, so it’s two mortgages really founded on the tech first approach to mortgage lending, as we know is one of the industries that has kind of remained archaic in terms of just providing the consumer experience, how the process works, and some of it is regulatory meeting. As a lender, we can make any make any changes. The government asks us to originate alone and in a certain way there is not much you can do about it.

I’ll give you an example. We created the record for fastest from an application to be ready for close in 6 hours, six and a half hours in the industry. But see, APB requires us not to close a loan for seven days after we take the application. So that’s. That’s what I’m saying. And some of the things that we are doing, we are kind of hampered by what the what the regulation brings. But other than that really is the mortgage approach is towards speed, towards making it more streamlined, user friendly. And we do that by a lot of technology implementation that we have done. Some of it grown in-house at instrument and some of it that we’ve used our partners to build.

Buck: One of the things that I understand that you guys are using a lot of is some of this artificial intelligence. Let’s let’s start with something basic, I think. But this is not necessarily a technology group, but, you know, maybe you want to start with like sort of the concept of what Shashi Petit is and, you know, kind of take take it from there and explain how some of this might apply to your business.

Shashank: Yeah. And anyway, I’ll probably back up even a little bit more and just just explain artificial intelligence. I know the audience is pretty smart, but one of the things that we see all the time and it’s this not just in the mortgage industry is is that a lot of it being is what we are hearing, especially people ask two or three years is that this is a driven we it everything seems to have the word A.I. in it. It’s like crypto or blockchain from a couple of years.

Buck: Yeah, I guess 100% right. And yeah.

Shashank: One thing I wanted to explain to the audience is that not everything is the AI, and just because they say it’s to write artificial intelligence essentially is is the is when machine is able to do some things that humans do, they’re able to recognize speech, they’re able to make decisions like humans do when something is programed to do something that’s just advanced programing, like if I had a chat bot and I programed it with every possible question that would ask when Buck would come to my website, that’s not artificial intelligence. Me look smart. Oh, I’m asking you, do I need a mortgage? And it knows the answer. I want to can be connected to a loan officer and it knows the answer, but it hasn’t been programed to do that, that we preempt the questions that you might possibly ask, given our experience that when users come to the Web site and ask these questions, that’s not artificial, since artificial intelligence is usually a huge amount of computing capacity, that’s required for a machine to learn by itself, even something as basic as if a machine were to recognize if this picture is of a cat or not, that will require it to be fed hundreds and thousands of photographs of cat for it to figure out a pattern saying this is how cats look. And on the other side also fed hundreds and thousands of pictures that may look like a cat, but it’s not a cat. Might be a baby lion or a baby tiger that has paws and whiskers and small, but they are not cats. And and then it’s able to figure it out through machine learning, through through hundreds of attrition. That’s not a cat. So that’s artificial intelligence. And I just wanted to throw that out as because search terms, we are like, oh, this is artificial intelligence or it’s AI driven, but it’s not always the idea.

Buck: Yeah. So really, I think like when I think of it, I maybe, you know, just for clarity, maybe you can correct me if I’m wrong, but it’s really like a learning system, right? It’s, it’s, it’s not just preprogramed information. It is literally sort of learning on the go. So it’s, you know, it’s going to take information that comes in. It’s going to take questions that come in. It’s going to learn from those things and potentially create more efficiencies because it can start, quote unquote, thinking rather than simply spitting out preprogramed information.

Shashank: Absolutely. And that’s why and that’s where I think GPT comes into us, as you asked us, that it’s a job to please is a truly AI model. And what’s revolutionary about it is really so far practically all artificial intelligence that I’ve seen before. Chad GPD has been about solving a specialized problem. So it could be about how to become a better writer. So it could do copywriting for you or to become better marketer, or you’re solving one problem or the other. And that’s why artificial intelligence was mostly being deployed, charging beauty for the first time at a very mass level really brings a generating by origin by meaning it didn’t plan sort of it at least aims to solve for hundreds and thousands of problems and not just one last I checked. So in the real estate industry, AI and chatbots like ChatGPT are having a significant impact. One area where they are being utilized is in customer service and support. Chatbots can handle customer inquiries, provide information about properties, and even assist with the mortgage application process. They can answer commonly asked questions and provide a more streamlined and efficient experience for potential homebuyers.

Additionally, AI is being used in data analysis and predictive modeling. By analyzing vast amounts of real estate data, AI algorithms can identify trends, patterns, and insights that can help real estate professionals make more informed decisions. This can include predicting property values, identifying investment opportunities, and assessing market conditions.

Furthermore, AI-powered virtual assistants are becoming more prevalent in the real estate industry. These assistants can help automate tasks such as scheduling property tours, sending notifications to clients, and even providing personalized property recommendations based on the client’s preferences and requirements.

Overall, AI and chatbots are revolutionizing the real estate industry by improving customer experiences, enabling data-driven decision-making, and automating various aspects of the real estate process.

Buck: It’s fascinating. So in your business, you mentioned that AI is streamlining the underwriting process for instant mortgages. Can you elaborate on what aspects of the underwriting process AI is helping with to reduce the timeline from 45 days to just a week?

Shashank: Absolutely. One of the things AI can do is bring efficiencies to the table. For instance, loan guidelines in the mortgage industry are extensive, similar to the complexity of the IRS tax code. It can take a significant amount of time for underwriters to sift through hundreds of pages of program guidelines to find specific information related to unique borrower situations. AI can analyze the guidelines and identify the relevant sections more efficiently. It can also pinpoint bottlenecks in the process, allowing for improvements to be made.

Furthermore, AI can assist in automating tasks and providing quick access to information. For example, I recently experimented with a Facebook language model that can be installed on a local device. By uploading relevant documents, such as research papers or industry-specific information, you can create your own repository of knowledge. You can then ask questions and receive answers based on that information, streamlining the information retrieval process.

Buck: That’s interesting. It reminds me of the open-source nature of blockchain technology. Do you see any similarities between AI and blockchain in terms of their potential and openness?

Shashank: There are some similarities, but they also have distinct differences. Blockchain technology has generated a lot of excitement in the past, and it’s still relevant. However, it has been closely associated with cryptocurrencies, which have faced some challenges. This has caused some people to shy away from blockchain.

But blockchain technology has applications beyond cryptocurrencies. For example, in the real estate industry, it could revolutionize the process of title transfer. Instead of going through county courts, titles could be securely held on a blockchain, and smart contracts could automate the transfer process upon receipt of payment.

While blockchain may have lost some favor recently, it doesn’t mean it has lost relevance. It’s just not receiving as much attention and investment at the moment. However, both AI and blockchain have the potential to transform industries and improve efficiency in different ways.

Buck: I guess the comparison I was trying to make is less about comparing them as the same type of thing, but more about looking at it as an open-source opportunity for people to develop.

Shashank: The interest in it is huge. By the way, if anyone wants to test it, go to GPGforall.io. Even if you have zero programming skills, you can hire someone on Upwork. I’m not a programmer, and I got someone to do it for me. It took him less than 30 minutes. The instructions are step by step, so even if you can’t do it, someone else can.

Buck: Let me write it down: GPT4all.

Shashank: Yes, “for” is the number 4.

Buck: Okay, dot IO. Fascinating. Let’s see what I can come up with.

Shashank: Very cool. So yeah, let’s talk a little bit about your business too. You guys are obviously doing really well. To be honest, I’d never heard of Instant Mortgage. Is this available in all states?

Shashank: We are in 28 states, so not all states yet. But those 28 states account for about 70-75% of all mortgages in the US.

Buck: Mostly in California, right?

Shashank: Yes, I’m here in Silicon Valley.

Buck: Right, got it. So, how does the process with Instant Mortgage compare to other mortgage companies?

Shashank: If you go to InstantMortgage.com, you will see that our entire smart application process is way more advanced than most of the industry. We start underwriting the moment you start entering the application. By the time you finish the application in your own borrower’s portal, you will already start seeing the documents we need and the signatures required. The loan is almost half underwritten by that point, which is probably the fastest you will see in the industry. We also have a user-friendly borrower’s portal that provides transparency on your loan’s progress. We have 100% transparency on where your loan is and what’s required from your end. You see all the loan dynamics at that point in time. The entire user experience is very advanced. We still have expert loan officers and processors in the back end, so it’s not just a technology solution. You get expert advice as well.

Buck: Is it easy to get pre-approvals with Instant Mortgage?

Shashank: We have something called a prequel where you don’t have to go through the entire process. In 90 seconds, you give us some basic information, and we run a soft check on your credit. It does not impact your credit score. Then we run the entire guideline underwriting engine and give you a prequel letter, which is the fastest in the industry. If you just want to understand how much you qualify for and you’re not ready to make an offer, you don’t even have to reach out to us.

Buck: So primarily residential, including jumbos and all that. Any limitations on the residential side?

Shashank: We also do commercial loans, but the process and user experience would be very different and take longer.

Buck: That is fascinating. One last question for you. Who is Rachel and how did you create her?

Shashank: Rachel is the mortgage industry’s voice-first digital human. We started with websites, then moved to chatbots, and now digital humans are the next evolution. Rachel is more advanced than Siri and Alexa because she has a face. You can interact with her through speech, video, or text. You can ask for jokes or talk about non-mortgage-related topics. We just launched her with Jeopardy, so she can answer Jeopardy questions too. It’s a pretty advanced concept in terms of user experience.

Buck: So the business is Instant Mortgage at instantmortgage.com. This is really fascinating stuff, Shashank. I appreciate you being on the Well Formula podcast today.

Shashank: Sure, glad to be here.

Buck: We’ll be right back.