Due diligence with AI, ML and data extraction

• AI, ML and data extraction applied to due diligence.
• A real estate and note-investing angle on diligence tech.

How automation is creeping into document-heavy diligence.

file that I will be able to upload in our system will generate a exception report and merge with title and now you have a complete due diligence report and exceptions that you can give to to the seller as always subscribe and click the notification bell on our YouTube channel be active on our Facebook group East Coast distressed note investing .

And be sure to go to our website for all your note investing tools, resources, and available assets.

If you have any questions, email us at tradedesk at jkpholdings.

com.

David Nathan here.

How are you, my man?

Hopefully all is well.

Things have been a little chaotic.

I apologize for the craziness we're going on right now, but we're running into some issues that are out of my I guess control for a minute.

Today we're going to be doing some little special.

I think what we've found in our space is that due diligence became not only requirement, of course, but a problem, some situation because not all the data is easy to access.

A lot of the stuff we're trying to get to is all over the place.

Yeah.

.

So in your experience, Nathan, what'd you start off with?

And what'd you start doing when you first got into due diligence?

How did you manually do a lot of stuff?

When I first started, again, this was back in the day when there was never any equity, right?

So there was, it was always negative equity.

So I had to know what the value of the house was.

That was like number one job.

Number two job is to find out if there are any outstanding liens, , anything else besides the first lien that I've just purchased.

So I would do, when I very first started, I would do calls to counties and I would ask about taxes and things like that.

I would do all kinds of things like that.

And then the further I went along, I met a guy named Alex.

And he told me about these title searches and I'm like, oh, really?

.

And at the very first, especially, I'm not even sure how many times I called Alex.

I'm like, okay, what does this mean?

And like, how do you, how do I read this?

And like, what is that?

When it says this on the report, what is that?

And I went over it with him and he was very patient with me and, and coached me through how to read these reports that he was, that he was producing.

And, and it was extremely helpful.

And now I don't call him nearly as much.

Every once in a while, I still get stumped.

And I'm like, I don't know what .

So I'll still call either Alex or somebody on his team and just say, so what is this line?

I've never seen that before.

Yeah, I think for a lot of us, it was a lot of information in the beginning.

And we were getting not only going after Zillow and going after Realtor, all these sites that gather data.

Then we go to the county records.

And then it became a point where it was a lot of information everywhere.

.

In the note space, we're lazy.

Let's be honest here.

We're extremely lazy people.

And I think what we've found is we like things to be as simple as possible for us to be doing things in a clean layout.

We're talking in just Wednesday about the idea of don't send us PDFs.

Send us something a little bit cleaner.

Spreadsheets.

In a spreadsheet, I can easily move it around.

it, that kind of thing.

It's just way easier.

But yeah, clean and easy to read.

I've seen total reports from some people where it takes me a while to just decipher it.

And so that's actually one thing I really liked about ProTitle.

And we'll have Alex come on in here in just a minute.

But I find their reports are really easy to read, very straightforward.

Those first couple of pages of the summary, and it's just all the information's right there.

.

Absolutely.

I think for us, we often want things to be as clean as possible for us to look at, but also in a layout that is consistent.

In our space, a lot of things aren't consistent.

When I first looked at ProTidal, I, oddly enough, easily understood it because of the way it laid out and things kind of brought to .

So it was really cool to see that happen.

When we come across stuff like this, it makes our job as investors a whole lot easier to do, which is great.

So when we say this stuff, we bring on someone like Alex to kind of share his knowledge and experience.

It's extremely beneficial, right?

So we want to bring on Alex, maybe screw him in here.

It may be, , which is great.

If not, let us know.

But we're recording this.

It'll be on YouTube real quick.

So, Alex, how long have you been doing this title stuff?

I feel like forever.

It feels like forever.

Hey, guys, great to be here.

Not only you're my clients, you're my friends, right?

Every time we meet, it's a small group of folks that know each other.

I love to hang out with you guys.

Next time, I'll see you at the conference.

.

But I've been doing this stuff since 2007 as ProTitle USA, but prior to that, I was the investor, just like you, Dave, and you, Nathan, into apartment buildings and student housing.

I used somebody else and I absolutely hated them.

And therefore, I said, listen, I want to build something better, which is easier to understand for the investors.

And I launched ProTitle, which was an experiment for me, but basically it was successful enough to make it a full-time business and it grew exponentially to the point of us today sitting talking to each other while we have contracts with some of the large funds out there, billion-dollar funds, GSCs, FDAC and others that now utilize ProTitle as one of the key vendors the data.

.

And we do service quite a bit of investors in the capital markets, both in residential and commercial space.

So that's sort of my story, how far we go back and how far we We went forward from the time we started.

I'm trying to remember when we met.

It was probably like 2012 or 13, somewhere in there, 14 maybe.

I'd say prior to that.

Somewhere in there.

I don't remember.

But I remember meeting Alex and going, man, this guy is looking to make a mark.

He wants to dominate this space.

And in a big way, I think you've done that.

.

You go and ask anybody, who do you get your title reports from?

Almost always, ProTitle, that's the name that comes up.

Yeah, I mean, I'm pretty happy that we dove into this space and I wanted to design the best product for the moment.

In fact, right now I own three companies.

So I'm a CEO of three companies.

ProTitle is one of them.

I have acquired a nationwide doc prep and recording firm.

is called DocSolution USA out of Houston in 2021.

And now we're growing like crazy.

And the company that we're probably going to touch on this call is called OneDiligence.

That's the software company that is targeted to solve the data extraction, machine learning, AI aspects of loan files, right, when you do diligence.

Combining all three creates a very powerful one-stop .

So the more I look at the market, the more I see that with a lot of automation and AI, you can consolidate a lot of the services into a single package and make it so efficient that not only the title, credit and compliance, the loan file review, critical document review, data extraction, .

That's all going to be solved in a single shop.

You don't need three vendors.

You just need one or at least one umbrella.

So that's the goal for me next to go ahead and create this single structure of seamless due diligence on the market that nobody else has.

Interesting.

Holy cow, man.

is something I talked about a lot, and I preach a lot, simply because it saves time, reduces errors, and allows you to do what you do best.

A lot of us are not researchers.

We're investors.

We're numbers people.

We crunch stuff.

For us to investigate and do our stuff, to be honest with you, it's not that we get lazy and it's like, well, I've done enough.

However, we enjoy the data that we're able to pull in.

, in a quick process, we like the data, we just don't like going to get it, right?

And the more data we have, the better we are at it.

But at the same time, I think for us, the idea of AI is something that I think is kind of a hot topic right now, and we don't fully understand it, right?

So how did you get into this idea of not only pulling title in O&E reports, but .

Yeah, I mean, let's try to focus on what AI is, right?

So a lot of people hear the buzzword AI and machine learning, and we really don't have a good understanding what it is, specifically for what we do, right?

It's a very hot topic.

Everybody's talking about it.

It's as hot as, you know, blockchain.

That's another hot topic, right?

.

Well, probably, I'd say 10 years ago, if somebody would say I'm working on AI, what I would think it's a bunch of killer robots running around, just like in James Cameron movie, right?

So that's AI.

It's somebody that, you know, is a human with a robot brain and does something that replaces the human.

.

So it's definitely something that a human creates to do something for a human, but it's very narrow focused, right?

So you hear the buzzwords, chat GPT, right?

So now it's very, what is that?

Well, it's a great tool to replace all the admins in all the companies that really is a tool to analyze the text data and, you know, continue the conversation to find out what the what the client wants uh or needs to do right so it's it's analysis based on some uh textual data the um there's things like um rpa right that was pretty hot and still very hot right so it's everybody probably saw in the airports a sign for ui path right very hot company that i think started in .

So for things that are robotic and automated or really mechanical, those things are great.

The price for that little robot is $10,000 a year, roughly, plus or minus, right?

So it's a robot that sits and does something, accounting or some tech stuff or some updates or some coding even.

That's not what we are here to discuss, right?

All of them are doing the task that human would train them to do.

The AI and machine learning in our mortgage space, it's really very simple.

The task is simple, but to implement it, it's very tough.

The task is I have a loan file where Nathan is buying a loan.

He has the loan file, which is 6,000 pages.

I would really hate to look at every page.

and find out what it is, right?

Because it's up to me to run diligence.

And I don't want to do it.

I'm too lazy, right?

So what I want to do is I want to take this 6,000-page document, load it somewhere, and get something back that tells me, hey, you know, you're missing a note.

Or it's out of compliance.

Or, you know, the title policy ensures for the amount which is too low, right?

Than the mortgage itself.

all those bells and whistles and whistles and whistles and whistles and whistles.

Ideally, you want to have a black box that automated everything, right?

So let's call that AI module for diligence.

That's what one diligence is.

It's a system or a platform that allows people to drop their loan file inside and get the exception report back, right?

So what is machine learning, right?

That's another hot buzzword that we .

So machine learning is for any document that machine doesn't understand.

From either understanding what this document is or what data is inside the document called data extraction, right?

Have an ability to learn on the fly what this document should be.

And then next time the document is a part of the loan file, it recognizes it.

which data to extract, and how to use that data, right?

So there's a lot of rules to understand.

All right, so this is a loan application, and this is specific for whatever it is, and I've never seen it before, but now I saw it, and I know how to learn it, and next time this application loan application bill comes up, I know exactly what it is, and besides that, you know, I know where to grab the data points for me to make a decision that this is for my property, for my borrower, for the right mortgage and so on, right?

Whether it's compliant, it has all the things that you need in the loan app.

And if I don't have it, then I will come up with some sort of an exception report that will flag the investor like Nathan to say, hey, I'm missing demographic form, right?

or it's not compliant.

So you would ask the seller for, hey, give me a collateral for this file again.

I'm missing this page or you have to choke up a discount, right?

Because now I'm facing a risk for a borrower saying that I didn't sign the right document.

Or maybe you have unsigned loan agreement or unsigned HUD-1 final statement, whatever it is, right?

So now you have to be able to make rules to check things on the fly, right?

And that's called machine learning, right?

.

So I know it's a long answer, but I hope that that sort of puts it in perspective.

What's AI?

What's AI for our field, right?

And what to do with it?

Well, I could send you a digital collateral file and say, okay, give me the report at the end, and then it'll give me like a one-page thing to say terms as well as anything missing, documents, , that kind of thing?

Yes.

So I don't know if it's going to be one page.

It might be a pretty comprehensive report.

If that's the limitation, it may be multiple pages, but I went from 100 pages down to 15?

Maybe, but it's probably what you get is, .

So for every loan, you have an exception that we would find.

But conceptually, that's absolutely correct.

If you give us a file, you say, I have no idea what's there.

Seller gave me this tape.

And then I want to make sure that what I'm buying, what I'm counting the seller for this tape is absolutely accurate with the loan files.

And there's no issues or discrepancies between the loan file of 6,000 pages and my loan page.

, right?

Because you're using the loan tape as a tool to bid on a loan.

And you're trusting the seller to make sure the loan tape is accurate, right?

Trust but verify.

And therefore, you have to have some tools to be able to see if you're buying the accurate loan, right?

And all documents are there.

And you don't have to chase the seller when you close on a transaction two years after that you need .

How long does this take a user that logged in and started doing this?

Well, we typically don't let the users play around with the system as far as QC-ing, but at the same time, we do provide the SaaS-like approach, or we provide our QC staff to go and look at the documents.

.

Someone asked, do you have a background in AI and technology?

No.

I am one of those people that, you know, I love to invest myself in the latest and greatest.

I wanted to prove to myself that I can do it.

And let me take a step back.

And this is an interesting topic, because if you are a software for engineer, or if you get yourself in IT or technology, and you start investigating on how to build a machine learning system, and you use the Google official guide on this is how you build the machine learning or AI system, I disagreed with it completely.

I disagreed with Google.

Throw stones at me.

Kill me.

That's it.

I just disagreed.

.

Google says that you have to have a human in the loop, which is a human to check that the machine extracted all the data points correctly.

So you have somebody probably offshore, a lot of people that, again, I've seen all the solutions out there on the market.

Trust me.

I looked at it.

I tried it.

I was not happy with any of it.

So I said, just like the title search, .

I tried those vendors out there, and one vendor cost me a $30,000 loss.

I said, I'm doing it myself.

So I built the system for a profile.

Same thing with OneDiligence.

I looked at all the solutions out there.

All of them have something called human renewal.

And I said, why do I need to bottleneck for my loan data extraction or data points extraction from the loan files by a human that sits somewhere in Philly or India or Vietnam, anywhere, that they're looking at the loan file without complete understanding with the probably exposure of private data and bottlenecking on their review to give me a data back, a crisp data back.

It makes sense that the human loop should be there, but to me it doesn't.

.

The reason why I say that human in the loop to me is not the correct approach is if the system is working perfectly, I don't need the human.

I want the data instantly.

So I loaded Nathan's loan file.

I want the data right away.

Even though it's 90% of data comes back to me and not 100%, I don't want to spend time on that human to verify that all my data is correct.

.

So human in the loop aspect to me was a waste of time.

It would bottleneck me from my system.

So I have designed the system with the help of very smart, genius programming guys.

One of the guys is my friend who lives in New Jersey, and we actually started ProTitle together.

But he joined on this venture to design this beautiful system.

that bypasses human loop and instantly delivers the data within, I'd say, minutes, right?

So the challenge with the system is you have to have two appended steps back to back.

One is called indexing.

What is the indexing?

Indexing is a step that's required to identify each document within the loan file, right?

So in other words, you know, I have 6,000 pages, and I wanted to identify a document within 6,000 pages, which is called a note, policy, mortgage, loan app, hot one, and so on, or appraisal, or a mod for that one.

Once I identified the document, I wanted to submit it for extraction, data extraction.

So I take this document, and I submit it for data extraction, and then, you know, a few minutes later, I get the data back, and I do something .

So all this process, I currently didn't spend any human labor.

I get the data, I run the rules engines, and then only now I get the expert human to run the QC on the whole system, right?

So in other words, now I have an expert person for crediting compliance or for document critical review or whatever the less mod data .

Right now, it's very popular, right?

So you have a lot of forbearance.

You have a lot of deferral agreements.

You have a lot of mods on the file.

And very frequently, the data from all of those documents don't match to the tape, to the servicer tape or to the investor tape because of last-minute agreements or last-minute changes before the signing of the mod during the COVID time or what have you, right?

.

And I'm getting too deep, I know.

After all the documents are signed, the P&I, or the payment and interest, does not calculate directly to the state of maturity.

And that's a normal, right?

So in other words, you have some deferral agreements with no interest-bearing amounts and some mods with different schedules.

, same as a data tape, it's a complete mess.

And now you don't know what you have, right?

So you have to have a rule engine to calculate that your amortization schedule is correct, right?

So all of those things can be done under the hood with some sort of rule check.

And let's say Nathan says, you know, calculate that my latest mod that I signed with a borrower and all the deferred agreements, deferral agreements and all the, you know, forbearance agreements and post-bankruptcy agreements, , whatever that may be, calculates correctly to the maturity.

And if it doesn't, I'm going back after the seller and say, hey, something's wrong.

You gave me the wrong tape.

I'm basing my investments on your tape, and it's completely, you know, bogus.

So those kind of things can go under the hood.

And coming back to Nathan's points, you know, am I giving Nathan, you know, 50 pages worth of report?

No, I'll just have one exception stating that.

for this loan, the amortization to stated maturity is above, let's say, whatever Nathan told me, $5,000 or $1,000.

And therefore, Nathan has tools to go and negotiate and make a decision.

Do I discount on this loan or do I go and just tell seller, you take it back?

For those people who are either newer or don't know the process, let me give you what .

, it costs $1.

50, $200.

00.

And we get an exception report in a few days.

This is doing it quicker, easier.

And a lot of attorneys really don't like to review cloud files.

It's not really worth their time and effort.

So that's an old-fashioned way of doing it.

And Alex is absolutely right.

It doesn't really need someone to understand what that looks like.

.

But you can connect things to things back and forth and see the chain being completed because it went from A to B to C to D as signatures.

But then you also have to figure out if the actual numbers actually match up, which means you have to run your own administration schedule, run your own data numbers.

Having all this in one is pretty impressive.

Yeah, that's interesting.

Very interesting.

Yeah.

I'm glad that I made people think.

We have some people here.

John Durville said that he does this for a living, I believe.

He's a background AI, and what you're doing is amazing.

It's very unique for the industry that has pretty archaic solutions out there.

So we'll just say that.

My goal is to basically throw money to break the market.

.

We're trying to change how things are done in a way that makes sense.

With the latest technologies, I think we should be able to create something that costs less, quicker, and easier to understand.

As a buyer and seller of the mortgages of real estate, you get a title with recorded docs, as David said, and collateral file of unrecorded stuff, right?

And typically, ProTitle would do a great job in understanding what's going on with the file, you know, from the recording docs perspective, liens, mortgages, taxes, and so on.

And we would generate a bunch of dashboards and analytical things just specifically for title.

And I always thought, you know, it's so weird.

We are just focusing on this narrow element of the due diligence.

without looking at the full scope of the diligence.

And that's what triggered me to go and buy the doc prep and recording firm, which now enables us to generate the assignments and releases instantly, instantly by OCRing and validating data on the pro title site, right?

So in other words, I already deal with the title site.

, I already get the mortgage and I get all the data points.

What prevents me from pushing a button and generating the assignments?

Nothing.

So why would somebody pay a separate vendor to prepare the assignment release?

To me, that never made sense to me because they do double work of verifying the assignment chain, verifying the mortgage data points.

It's already done.

.

Yeah, the machine learning and OCR function within the title report, now I have all the data, which is called cleansed, that can be passed to prepare the next assignment or subordination agreements or release of the mortgage instantly.

So that's what we've been able to accomplish.

And by the way, I also carry the custodian capability as well.

So if you need to store the files, we have a huge warehouse with the video surveillance .

But that's one part.

And the next part is the collateral file that I'll be able to upload in our system will generate an exception report and merge with title.

And now you have a complete due diligence report and exceptions that you can give to the seller.

Now, the biggest problem on this market, is cost, right?

So in other words, with COVID and raise of the employee wages, the cost for diligence went up probably doubled, right?

If you're looking at the collateral reviews or compliance reviews.

So what we're trying to do is find the most cost-effective solution out there.

by not using a fancy buzzword technologies in AI such as, you know, say, neuro video card graphic capture of image on the page and figuring things out.

Now, that's too expensive, by the way.

So if I would be processing photographs, that would be perfect technology, right?

If I would be processing text, all I need is The NLP technology, which is the natural language processing, understanding of documents, right?

Or standard form document approach where if the form itself doesn't change, then I can really inexpensive grab the data from there.

That's the other characteristic of a node investor.

A, we're lazy, B, we're cheap.

So if we're going to pay for due diligence, .

We'll skip it.

We'll look it up our own county record, right?

That's what we typically do.

So when you say this is going to be inexpensive, what is that?

I mean, what do you think that will be around if you can?

Just to scrape the file, let's say to produce the critical document review.

for each collateral file.

And those are usual suspects, right?

So it's the security instrument, it's a note, it's a launch, it's HUD-1, and so on.

Probably we're talking about anywhere from $40 to $50, maybe even $35.

So that's cheap.

.

And that will produce for you the exception report that will tell you what document is missing.

So that's perfect for lazy investors.

If you have a, let's say a 3,000 to 5,000 page file, drop it and get the report.

You're done.

When you talk about credit and compliance, probably, you know, for a non-volume clients, we're probably looking at, you know, 170-ish.

.

And that's complete credit and compliance review with, of course, no compliances because that's a separate fee.

Interesting.

So what about in the case, let's say it's the opposite issue where it's not a 200, 300-page document or 6,000 or whatever.

Let's say it's a 10-page document where it's a seller finance deal.

And the indexing in that is going to be different.

They might not even call it a node.

So how do we...

.

Does that work still?

It's very niche market.

I'm focused on the mainstream markets.

Let me take a step back and give you a few things to know as an investor of what we need to train our system.

I always thought that if we're big enough as a company, you have to have your own training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training is that field to be extracting, which is a new field.

And the vendor would say, well, stand in line.

It takes six months to do.

I said, all right, I don't have six months.

I need it now.

So it's easier for me to build a system and build a training system for data attraction, build a training system for the indexing.

So for indexer to work, we need 200 samples.

.

It doesn't matter which documents, right?

Seller finance documents or contract for deeds.

I know it's very, very niche.

We need 200 samples of single type of documents for a system to know that this is a seller finance or this is a contract for deed.

And it'll always, whenever it sees that type of document, it'll know, okay, I know what that is.

And we will use the same set of 200 documents to extract the data .

Right now, our production is about, we can train up to 20 new documents a week for the indexer, and we can get two new documents a week for the data extract, which means that...

What does a new document look like?

Let's say Nathan brought up seller finance.

I don't remember what it looks like, Nathan, even though I've been around.

Maybe I do.

.

I don't know if this is a unique document in our library, but let's assume it's a new document that we've never seen before.

Our system never seen before.

We always label it as unknown.

So Nathan said, all right, I really have this ongoing business where I need you to recognize that document.

And you will send me 200 samples under NDA for our system to train on those documents, right?

.

So we cannot use the whole loan file.

Just give us, say, page one to page end of that document and no blank pages and no dark pages, just that pure document.

So the system, like NLP, would understand, read the text, understand what this document is, look at the header, look at how it's structured, and really understand what this document is.

And after that, we'll take that and say, Nathan, what data fields you need from seller finance notes?

And you'll say, I need this.

.

I need borrower name, lender name, and all of the good stuff, and I want to know if it's signed or not, and if it's not, and so on.

We'll take that, and next time, we already know what this document is, and what data points to extract.

That's what I mean by new document.

Yeah, that's so hard, because I know Nathan and I went through this process, and that's going to be a struggle, I'm sure, for you, because what they call borrower, I think I had 15 different variables.

and what they call borrower.

And just the different type cases were just unreal.

But at the same time, there's also 3,000 counties nationwide.

And how hard is it to get all 3,000 counties?

You've already accomplished that.

So I challenge you, but I'm afraid to challenge you because I know you'll probably figure it out, which is tremendous.

So how quickly does this typical process, if I'm going to send you a file that you've seen before, .

We've got the cost.

How quickly can this expect it to be returned to an email back to us with a document inside of it?

Yeah.

So how quickly?

Well, let's say we're doing the critical document review for the regular mortgage.

Again, I don't know what you're scaring me with the seller finance.

But the regular mortgage probably will be, you know, a couple of days because I want to allocate some time for the human.

to get in and to analyze the data, right?

As far as the machine and AI portion, I'd say probably 20, 25 minutes.

Oh, amazing.

Right.

So could you say that people who are typically sending this stuff to collateral companies to check, attorneys to check, are they still needing to do that?

Or is that something that they just need to do now?

And you, as you grow, you won't have to, or are you- No, I, .

I'm not here to replace attorney or attorney opinion.

No, no, no.

So we don't provide legal opinion, right?

So that's the disclosure.

I'll give you an example.

Let's say you're buying a note in New York and you don't know what anti-Engle law is that was signed by a governor at the end of the year, which really makes your notes unenforceable.

Attorney would know.

An attorney would know , how to check, what to check from the perspective of prior foreclosure action, the formation agreements, releases, and so on.

You don't.

Attorney does.

We are not the attorneys.

We're not going to assume the liability or interpretation of the documents.

We provide the data and the data points.

And I think in different states that there's different things that you need to know.

We don't.

.

So we analyze documents.

We analyze the standard compliance, which is, I guess, federal and state compliance.

We analyze the title data.

We're experts in it, right?

To interpret the data, if you're not an expert investor, then you probably need an attorney.

That's awesome.

Then we can take that report that we get back, and I would probably still send it with the original collateral file over to the attorney, but with that report attached saying, we've already gone through it, .

Here's the summary.

Is there anything we need to look out for?

Is there any weird gotchas in here that we're not prepared for?

Exactly.

It's not only you send the original collateral, we bookmark it for you automatically.

So that's a part of the machine process where once you index the file, we return the bookmark file in the PDF bookmark to you and to the attorney.

the exception report, which you can send to not only to your attorney to review, but to the seller and say, can you comment on all of those exceptions that, you know, one diligence pulled up and therefore, you know, maybe their attorney will look at it and comments on your exceptions by saving you some dollars, right, to go to your attorney.

.

I think for a lot of people, this is up here, just below over their head.

But what they're going to find out shortly is that this is going to save them time, energy, be able to respond to sellers quicker, and be able to just analyze things without guessing at things, which is a really difficult thing that people do regularly, is they guess of what they think it is.

.

That's a real dangerous thing to do.

What do you say to those investors who are getting into this space, what should they expect and do with the information versus, you know, because too many people rely on easiness, right?

Should they be diving in and learning a little bit more about this stuff before using this, or should they dive into it and learn it afterwards?

That's a tough question.

.

So I always say that's, you know, I'm in the same spot, by the way.

I'm the investor myself in non-real estate related things that I have no idea what to do, how to do, what to look for, how to run diligence.

And I rely on somebody with expertise to say, hey, you know, I'll give you an example.

I invested in oil and gas funds or fracking wells.

And so I've never dealt with it, but sounded logical, right?

.

And somebody guided me through that investment.

Same as buying notes or self-finance notes.

If you have money and you say, hey, I want to invest here without knowledge, that's a little bit dangerous, right?

So I would follow advice from David and Nathan on what to do, how to do, and go through the waterfall approach.

Otherwise, you're going to lose money.

Yeah, that's my advice to the investor.

Be careful.

Selfless plug is, , you know, which I don't make any money on.

It's out there in Amazon.

Buy a book that I put together for the title, and that at least gives you a basic understanding on how to read the title reports.

That's part one.

I don't have any book on diligence or doc prep, so I'm too lazy to write it.

It's just I don't have any time.

Maybe I need to hire a ghostwriter.

But, yeah, but in essence, you really have to get education, right?

.

Actually, on compliance, there's a few free courses, I think, that are being offered by our governments.

So that I've seen by MBA.

I've seen if you're a registered member of Mortgage Banker Association, they offer some free courses from the investment perspective on credit compliance or dot prep and things like that.

So get education or get together with somebody who knows the business.

Otherwise, it's going to be a tough time for the , a new investor.

So I did want to pop in before we disconnect and Nathan asks his question.

We're getting top of the hour here.

One of my everyone that we actually are for DME videos, we've recorded them.

Unfortunately, Alex was not there.

However, I would encourage you guys, if you're looking for additional trainings, understanding and everything else, we do have the recordings of both days at the DME.

No conference that happened early.

.

I'll put it into the YouTube channel as well.

Please go ahead and click on that.

Take a look at it.

You can purchase both days and you'll see all the recordings with speakers and whatnot.

So I'll let Alex, I'll let Nathan ask Alex our famous last question.

Yeah, what do you see on the horizon, Alex?

Like you've been doing this a while now and you've got to, .

A unique perspective on where things have been and where they're going because you're looking at really the past.

What do you see coming up in our future?

What's your housing projection?

Housing?

Let's see.

This is not my projection.

It's more from Mortgage Banker Association.

I think David from Corny is the data scientist for MBA and I attended Texas mortgage banker conference where he predicted that all the COVID money and all the PPP money and all the government supported money that was given out in the past couple of years for 50% of folks will run out by September.

So he shows on the curve that after September, that are really in trouble, in distress, will probably be forced to sell, right?

Or they will get into distress.

In fact, it's the banking crisis was avoided by FedNow policy, where all the banks that tapped into $8 billion worth of Lifeline to save themselves, from being retaken by FDIC.

However, it doesn't solve their problems, right?

If the rates will stay low, sorry, if the rates will stay high, they're still in trouble.

And I don't know how much they can borrow.

So commercial real estate loans that were originated by those smaller banks, those loans will drive those banks down and I don't know if they'll survive.

So it's yet to be seen.

I feel that a lot of things will be unpacked towards the end of the year, Q3, Q4, and you'll see a pickup of foreclosures and distressed markets after September as far as the number of files.

So that's what I see going in the future by looking at the data from MBA, by talking to my colleagues and funds that really, .

We have a lot of dry powder in the market right now.

Billions of dollars are waiting to be invested in distressed assets, and they're waiting for the right time.

So if you are a new investor and you're getting into this market, I think it's the right time, right?

You get educated, you try one or two.

Yeah.

Third and fourth quarter is that time where we really be looking into what we should expect.

And we actually had the NBA come to DME, or let me phrase that.

Nathan had the NBA come to DME.

and was one of the open speakers and shared some of the charts we talked about.

And it was shocking.

Yeah.

Not what I expected.

Fascinating information.

She shared some really great stuff.

Yeah, this is, I think this is a great time to get in.

It's never really a wrong time to get in, but this is a great time to get in because there's more coming.

We want to respect your time, I'm in.

I will disconnect in the live feed.