Tidalwave

Transcription:
Transcripts are generated using a combination of speech recognition software and human transcribers, and may contain errors. Please check the corresponding audio for the authoritative record.

Diane Yu (00:08):
Hello everyone. My name's Diane Yu. I'm the founder and CEO of Tidal Wave. Tidal Wave is in the market offering an AI-powered mortgage point of sale, also known as a POS, plus our AI agent acting as a loan officer assistant. Working with the loan officer, it helps engage the borrower, collecting the secure permission, evaluating the document, and creating the mortgage application so that we work hard and your loan officer doesn't have to. That's our promise and the results for the deal. I'll have Chris here running the demo. The team does everything on my behalf. Please, Chris.

Chris Olsen (01:01):
Thank you all. Pleasure to be here. What we're going to look at is a couple of different scenes. Right now I have my loan officer hat on. We can get a loan into the system many different ways, but my favorite way right now is through creating that lead through a conversation. We can call the borrower; we leverage Twilio to record the phone conversation, take that transcript, feed it into our mortgage contextualizer, and pre-populate the application to the borrower. That's really hard to show in a seven-minute demo. So instead, I'm going to paste a conversation in. This can be an email conversation that you might've had with your borrower. We are mobile responsive as well as have a native mobile app that this can be done through. We can see here, this is between Chris and John Homeowner. A lot of you are familiar with John Homeowner and probably have lent to him before.

(02:00):
So he is looking in Nashville, Tennessee—conveniently, I know a lot about that area—for around $500,000. It might be a mobile home these days. A common down payment is 20%. Does that sound right? Why, yes, it does. So we can take all of this unstructured data and feed it into Solo. Solo is to Tidal Wave as Alexa is to Amazon; think of it that way. We then save the source data for compliance so we can always look back at that. It is fast, real-time, with no human in the loop for any of the demonstration or AI technology that we have. We prefilled this application, giving the loan officer the ability to make any changes they may have both here and before we send it. From this, we're going to create a loan record. I can invite my borrower to this record, and now the borrower has the ability to complete that loan application with all of the data. We've all seen loan applications, so I'm going to highlight more of our agentic AI functionality today than the functionality for the standard application. We can take a full application leading to an AUS approval, running dual AUS. With that, let's log in with our borrower here.

(03:49):
The borrower also has the ability to log in with Google or Facebook. In this case, I chose a password because I reuse that application many times. The borrower also has Solo here to interact with. "What can I expect after I submit this application?" This is going to keep it within the context of a mortgage. We can also add lender-specific guidelines. So while this is a more general answer, we can be more specific to the lender. And then my favorite question—

Diane Yu (04:38):
So the idea here is the borrower can ask any question, having AI acting as the loan officer assistant to help with the answer.

Chris Olsen (04:49):
And always keeping it in the context of mortgage. So we're not going to get outside or run down a rabbit trail, search the internet, and maybe be talked to by another lender who's probably buying those mortgage-specific keywords. You're keeping your borrower here. I have credit, income, and assets. We've run pricing in the background and now upon submission, we're running dual AUS. While dual AUS is running in the background—I didn't use any first-party data, which we could—I'm going to upload some documents such as some W-2s and maybe my pay stubs.

(05:34):
While these are run, we're leveraging multimodal AI to take the data. It looks to see if it is the right document that was just uploaded and if there are any issues with that document. Here we can see that they can't verify the name; the name on the document doesn't match, or the date is outside of the range for the document I just requested. With the last little bit of time, I'm going to flow into the LO portal and look at this loan after it's come in and our agent suite has processed those documents along with some of the data. I've keyed up a loan here just a little bit ago, so we can see from our document screening agent that the analysis has happened. This is a purchase loan, picking out the product and knowing what's needed to submit for underwriting. We're giving the loan officer a quick list of what's needed with this loan before they should submit it. Also, any document errors will be highlighted. We've also scanned their bank statement, which we'll take a look at.

(06:51):
Our document screening has passed off the bank statement to our bank statement agent so that as the specific documents are identified, they go to the agent that is specific to that task. Here we can see I've surfaced any "buy now, pay later" transactions and some recurring deposits. We found additional income that can be used for qualification as well as additional rent and mortgage payments that the lender may have. We also give the ability to ask any questions around the transaction to the AI. If the loan officer is interested to learn more about that large deposit or wants the AI to write that letter of explanation to turn into a PDF and deliver to the borrower, those are all tasks that can be accomplished within Solo. The platform is not just for loan officers; it's for loan officer assistants, loan processors, closers, and underwriters. We do handle disclosures as well as hybrid eClosings through the platform.

(08:00):
All in one place. And I'm going to go back, as our time is running down, to the loan officer who might be interested to see some questions they can ask. I'll lower it and again, while it's easier to show on the big screen here, we are mobile responsive and do offer a native mobile app for interaction through the Tidal Wave platform. With Solo, we can see it's thinking. It's not just spinning out a response. It's looking at both the loan data and extracted document data that's now part of its profile, as well as having the ability to give general loan data.