Tea Mortgage Podcast 29 – The state of Artificial Intelligence in the mortgage industry

Podcast Transcription

AI’s Growing Role in Mortgages

0:03 – 0:12 Norman Schriever

Hey, everyone, what’s going on? It’s the Tea & Mortgage Podcast, episode number 29 with your fearless host, John Coleman. John, how are you?

0:12 – 0:22 John Coleman

Thanks, Norm. You’ve bullied me into doing this on video again today, so I had to put on my hot shirts and all that stuff. But yeah, hopefully, we’ll provide some good information for everyone today.

0:23 – 0:38 Norman Schriever

Absolutely. Absolutely. We covered this topic only about seven months ago, and we were trying to, before the podcast, figure out if it was a year ago, if it was three months ago, but it comes around quickly, and that’s the topic of AI and mortgages.

0:39 – 2:33 John Coleman

Yeah, well, I think in the last year or 18 months, this will be our third time discussing this particular topic, and it’s not because we’ve got nothing else to talk about. It’s more a case of, I don’t know about you, but I just feel the world is changing at a rate of knots. It’s exciting and scary, all at the same time, to wonder where are we all going to be in two years?

As you know, my philosophy has always been technology driven, and I’m embracing it as best I can. But with a smidgen of fear of wonder, am I going to bring myself out of a job at some point? But for the moment, it’s really, really exciting.

The speed of change, the speed of where things that were just not possible, things I would have asked my freelance developers, oh, can I do this two years ago, and one of the guys came to me 18 months ago and said, John, I think we can do this now. And then on every almost three months, actually, the thing you want now is now doable. So it’s absolutely scary to the point that I think I was telling you that I’m having a conversation with this freelancer, and I was kind of hypothetically saying to him, at what point would a human being basically start asking ChatGPT or whatever mechanism for health advice?

And this was nowhere to turn around and say to me, well, actually, currently, he was building an AI system for a vet, who basically would get his clients to go online to his system, talk about their pet or animal, and the ChatGPT will be diagnosing the pet. So that’s the bit that really scared me, by the way.

AI in Healthcare & Mortgage Processes

2:33 – 3:08 Norman Schiever

Well, even within health, now a big thing for people is to get their blood work done and all their markers, what’s your testosterone, what’s your blah, blah, blah. And there’s a health workup of like three, four pages of all these things. And AI could optimize and say, OK, take this vitamin, do this, you need to improve with this supplement.

And so, yeah, there’s a lot to it. And to your point, you’re saying, oh, AI is going to take our job. And I heard someone say it best.

They said, AI is not going to take your job. The person who understands and uses AI is going to take your job. So I think that’s where we’re at now.

3:09 – 4:56 John Coleman

And that’s ultimately, I know I kind of said that flippantly, but sort of as a fear that certain people have. I am very much of a view that what’s really my particular industry and any walk of life, a system like AI is going to get you to a point. It’s the little, it’s the differences, it’s the nuances, it’s understanding.

If certain things change and human empathy as well, words of encouragement, words of talking to people, when bad things happen, our banks have asked for different. So that’s never going to go away. So it is just a case of, as you say, whoever embraces this will become the ultimate customer-facing operation whereby everything is slick, but every customer, and this is what I’m ultimately striving for, every customer feels, my God, I must be their only client.

They’re nearly available for me 24 hours a day. That’s the sense of journey you want to bring people on, because buying a house is stressful, right? And the reason people don’t go directly to the bank, I think about 50% would go back, I think 50% use people like myself, but they want someone on their side, right?

I don’t think that’s ever going to change, from what I’m just being honest in terms of my analysis, because even if they’re great AI systems, each bank may have its own system, but they’re not going to tell you about all the other banks. So you still need someone to understand the nuances of all the different banks, but I’m fully, I get a huge amount of joy in every kind of incremental improvement in the service that we can offer our clients.

4:56 – 5:42 Norman Schiever

It’s interesting because, like you started, it was sort of like, oh, AI is a switch and you flip it and everything takes over. And that couldn’t be furthest from the truth.

It’s just little by little, this tool, this enhancement, this time saving, something resource for your process. And so it’s little by little integrating these things. And there’s places AI doesn’t work or it doesn’t really help.

But the best way that I’ve heard it as well, and I use this with a lot of my financial advisor clients, is AI is an amazing caddy, but it’s not the golfer. You know, and that’s the human element. You still have to be the golfer, but it makes the best caddy in the world. You know?

5:42 – 5:50 John Coleman

Yeah, that’s a new one. I like this one. I’m going to write that down and use it as my own. I hope you don’t mind.

5:51 – 6:15 Norman Schiever

No, please do. But yeah. And tell us a little, you know, about some of the advancements and some of the things you’re implementing in your business, too.

I think that the key is that you’re doing these things for a better client journey, a smoother process, a faster process. So you have more time to invest personally and give the human touch and the human element to each client.

6:15 – 7:52 John Coleman

Yeah. Well, I think we definitely mentioned this previously, and this was kind of the start of the AI journey, my technology journey, because that started beforehand in terms of just making it more efficient. But the biggest blockage in this process for anybody is the paperwork, right?

And then one of my freelancers came to me, John, at a seminar, and I’d always been talking to him, is there any way we can help with the system? He said, John, I think this can be done. So we spent a year building the system whereby when my clients upload the paperwork, AI basically checks it and will confirm for 90% of the paperwork if it’s right or if it’s wrong.

Okay? That then did set back, like, if I was honest, it’s working better for us at this point than it is working for the clients, and I’ll explain that in a second, because there’s still a level of technical understanding that’s required. For us, it means we’re getting the information as to what might be missing in a quicker way so we can feed that back to the clients.

Now, we are giving them the message of what’s missing, but clients being clients are sometimes not always seeing that message or looking at it, and we’re not delivering it, this is on us, we’re not delivering that message to them in the best way or the easiest user, what’s that, UI, is that an experience, the user experience or the UI, whatever that word is.

7:52 – 7:57 Norman Schiever

Yeah, I think it’s UX or CX, client experience or something like that, yeah.

Efficiency & AI’s Competitive Edge

7:58 – 9:47 John Coleman

Well, we’re not, so what I’ve built so far has made us more efficient, but hasn’t probably taken away slight the pain point for the client, okay, so it’s still resulting in a better service because they’re getting information quicker, but I’m a perfectionist here, I want, like what I started when I wanted this was very much a real-time system whereby when the client’s uploading their paperwork, they will be getting a sort of a yes, that’s brilliant, no, but here’s what’s wrong, and they’re getting it, but it’s not, we haven’t been able to deliver it in the best way, right, and I’m just putting my hands up to that, just from what I’ve seen, what I’m now currently building, Norman, it’s kind of sighting-scared, I’m trying to now build, get your head around this, it’s crazy, an AI avatar, right, so if the avatar will basically, the client, if they’ve got a question about, okay, well, I can’t upload my page because of the password, and the system won’t take it, right, well, the AI will answer in a video format, showing them what to do, okay, so it’s like having your own personal assistant there with you in the moment, right, for me, that would just lead to so much speed, but less frustration with paperwork, because it is frustrating, okay, and then what I’m looking at doing is building a kind of a, when they’ve uploaded all their paperwork, and say, some of the documents need to be manually checked, because that does happen with the odd one, because you can’t build a system that captures everything, you can build a system that captures about 90%, but then having a video at the end of the day, saying, we’ve checked your paperwork, here’s what’s missing, and here’s a video of what you need to do, right, and that’s not there yet, but apparently…

9:47 – 9:49 Norman Schiever

And it’s darn close, though, it’s darn close, right?

9:49 – 10:26 John Coleman

So I have someone at the moment working on a proof of concept on it, the second piece could definitely be done, the first piece is, I’m trying to make a statement, it could take me three to six months to get there, but what I really just want is a system where people’s paperwork isn’t the biggest pain, and they’re still, you can never take that away, because this would still annoy me, some of the banks in 2025, with their statements, have to be sent by the post. When you think about that, it’s just mind-blowing, right?

10:26 – 10:35 Norman Schiever

It’s the definition of inefficiency, and, you know, opening the window or the door for errors, or all sorts of things that can go wrong, right?

10:35 – 11:24 John Coleman

Here’s what happens on a frequent basis, someone will be like, can we get a statement? And the statement will be told, oh yeah, that’ll be in the post, you’ll get it in the next five to 10 business days, right? Then nobody, they don’t, the statement doesn’t come, right?

So then they have to ring again. And then when the time it does come, a lot of the other paperwork has gone out of date, because suddenly we’re into another month on new payslips, new bank statements. So, yeah, so as I said, I’m building a system to try to make it as easy as possible.

I’m still not going to ever claim that we’re building a system going to become, there’s still, I’m sure with time, some of the banks will improve their own personal technology processes to help in the whole situation, but for the moment, all I can do is my own bit, you know?

11:24 – 12:01 Norman Schiever

One of my financial planners, you know, talking about the banks, one of my financial planner clients, he talks about the technology adoption and AI for most financial advisory firms. And he says, they’re almost like mocking them. He says, they’re bringing you tomorrow’s technology.

No, no, they’re bringing you yesterday’s technology tomorrow. Meaning, right, they’ll get there, but it’s, you know, too little, too late, too slow. And it’s outdated.

So the banks will never, you know, outdo what you’re putting together. And that’s a benefit to your clients.

12:01 – 12:58 John Coleman

Well, it’s absolutely a benefit to the client, but it’s simple. We’re still, our paymasters to a point are still the banks, right? And there is a bank that’s come out there this year, a new bank has come into the market at the end of last year called Nua, and they are technically absolutely incredible, right?

Their systems are so slick. They’re sometimes getting approvals within two hours of an app;ocatopm. So it’s, so they’ve taken their app, their technology and belief in technology to the end degree.

And it’s just, it’s an example of what can be. But obviously the bigger institutions, well, they’re huge institutions. So it’s like trying to move an oil tanker.

It moves at a very, very slow rate, shall we say. So, and because there’s lots of expense in trying to change systems. And there’s probably lots of add-on things that you and I wouldn’t think of, but they need to be looking at the whole thing.

AI in International Mortgage Markets

12:58 – 13:24 Norman Schiever

Yeah, they have to, they can’t use tools, they have to build tools. They can’t just, you know, sign up for this AI service that works great, or bring on an AI consultant or someone to build a custom tool. They have to build a whole massive system for their whole bank, you know, and all their clients and because of compliance and everything like that and security.

And so it’s a, it’s heavy lifting whenever they implement something. But that’s why people should use a mortgage broker like yourself.

13:25 – 15:09 John Coleman

I agree. And you always do hear the bank sources are saying, yeah, we’re going to be, I’m not sure that the API, for example, is a term where you can get one system to be able to talk to another. And it keeps hearing banks, oh, yeah, we’re going to have this, our API system working soon.

I’ve been hearing that for over a year with some of the banks, right? So their technology change on moving is very, very, very slow. From my business point of view, it’s good to a point. But I’d still prefer if they were giving a better service, because my view on this is, well, I don’t view other brokers as my competition, right? Because I believe that the broker industry, if the service of the broker industry was all at the same level, right? Then why would anyone go anywhere near a bank to start with, when they have a broker who basically can give them the choice of every bank and give them tailored, specific custom advice that no bank will ever be able to give, because they only speak for themselves.

So I would really like if the banks were all at the right level from a technology point of view, and the brokers were all at the same level for the technology point of view. So then you have a much better system for everybody, right? And then suddenly you have, like, I’m not greedy, I believe this business for all the brokers out there.

So I’m not looking to dominate the brokerage. We’re setting my own standards, we’re setting our own standards, and we live and die by our own standards. But if we can kind of, what’s that phrase, a rising tide lifts all, and sort of…

15:10 – 15:10 Norman Schiever

Lifts all boats, yeah.

15:10 – 15:40 John Coleman

Yeah, lifts all boats. So that would be my view. And I’m pushing this the whole time.

And any other broker friends I meet with, I speak with probably boring them to tears, but I’m talking about all the stuff I’m doing in the background, I’m trying to help them. Some are more interested, and others are kind of, well, I’ve been doing it this way for the last 15 years, I’m going to continue to do it that way. So there’s always your kind of, your early adopters, your kind of people who come on. So then you’ve got the people who are just…

15:41 – 15:44 Norman Schiever

Some will, some never will, and some will eventually, right?

15:44 – 17:24 John Coleman

Yeah. But I think there’s a great advantage for my business in being an early adopter, you know, so I will continue to embrace it. And there was one other thing that I’m not sure where, just from a time point of view, in terms of, like, we would obviously have a conversation with a client, get a full understanding of where they’re at.

And then all the banks have their own different calculators, and we would basically have to put the information in and then create a report that then goes out to the bank. Now, that would normally be taking us a good two days to go through it all to make sure that roadmap that I believe in, that is the core of our offering, basically, we can now get that done within one day, sometimes on the back of this, right? And then ultimately, where we were, when someone is ready to go to the bank, we were probably taking four hours to analyse everything in terms of their, say, basics, their salaries, or some of the bank, all their bank accounts, we can now get that done within an hour. Right? So the time saving is just, it’s phenomenal. And I can see that only getting better and better.

So that’s to the point whereby the level of work we need to do, because again, back to the nuances, if you can get AI to do 85% of the heavy lifting, right, then you come in with your area of expertise, knowing the nuances, knowing the little differences that can give absolutely quality advice to clients in a much quicker timeframe. And that’s ultimately what I’m trying to achieve there, you know.

17:25 – 18:18 Norman Schiever

That’s a great point. The other thing I was thinking about knowing your business, and when you were talking, Ireland is very international, and you have so many expats or immigrants who are moving here and buying, and that means a lot of different languages. And so, and people who aren’t from even if they speak English, you’re not familiar with the Irish system.

So anything you could do with education, like you’re saying these interactive videos that, you know, it’s almost like a chat responder with 99% of the questions someone might have, but in a video form, well, with AI, you could instantly change that to, you know, a host of languages. And so it’s not just, you know, if someone is from India, if they’re from South Africa, well, South Africa is English, I think, but, you know, Brazil, India would be our two biggest non-European.

18:19 – 19:46 John Coleman

So we have a huge opportunity, we have a team designated by Brazilians working with me, we just focus completely on the Brazilian market. But India then is another big segment.

And we are obliged from a compliance point of view, for any non-mother speaking English speakers, we’re obliged to basically make sure that we ask them the question, do you have someone with you if you want, because a mortgage is a huge thing. And it’s called as a vulnerable customer. And obviously, in most cases, a vulnerable customer would suggest that maybe there was a death in the family or a serious illness or something like that.

But also, when you’re trying to do something big in a language that isn’t your own, do you have someone that speaks the language that you want there? Now, we have from the Brazilian side, we obviously have native speakers, so we can translate, but we are then obliged. And this is where AI is a great assistance.

Any advice that we give in Portuguese, from a compliance point of view, we also then have to have a version of that in English. So obviously using a system to take that piece and put it in. So that the advice in Portuguese and English, which if we didn’t have that, you can only imagine that that would be, we’d have to give advice in English because it’s all the advice that needs to be recorded in English, that would cause some time issues, shall we say.

Future of AI in Banking & Mortgage Lending

19:47 – 20:47 Norman Schiever

The other interesting thing is, and this isn’t necessarily you and your business, but in general, the mortgage industry, any savings in time, efficiency, cost through the whole pipeline with the banks, with the loan processing, with everything like that, eventually could be a benefit or a cost savings to the consumer. So some things like whether it’s the United States, they’re using AI a lot, they’re starting banks or mortgage lenders a little bit more. But for instance, there was a report that found that with AI, US lenders could reduce the loan processing by up to 40% through automation.

And also underwriting was sped up and less costly. And also, a big part of it was fraud detection and compliance. And so when you have these all these little stop gaps within their system with AI to detect certain things, all of that you add it up.

And if it saves them, you know, 20% manpower in time, and they could pass that on to the client, well, it only becomes more competitive.

20:47 – 21:31 John Coleman

Hold me a cynic. I’m not sure the banks would pass those savings on to the customer. I think that would go on to their, on to their bottom line.

And I think choosing your balance will be the right view on this one that would be if they can produce a better, quicker service. That’s a win, then you’re then down to kind of what each bank is going to offer different, different rates, and all the positioning themselves accordingly, right? So, but, yeah, in an ideal world, everything you say would be it would, the benefits would trickle down. I’m slightly cynical to think that the banks would have, you don’t think banks are, you know, out for the good of humanity?

21:32 – 21:41 Norman Schiever

What are you talking about, john? Well, think about it like this.

You always say competition is good, right? And if one bank rolls out some efficiencies through AI, all the others are going to take take notice.

21:42 – 22:00 John Coleman

And, you know, and there are probably already are like the bank that has come to the market would be probably that they wouldn’t be that they’d be more expensive on a rate point of view, right? But yeah, they’re lending, they’re going to shake up the market to a degree. I think I’ve mentioned that revolutionary in theory due to come to the market.

 

22:00 – 22:00 Norman Schiever

Yeah.

22:01 – 22:40 John Coleman

Later in the year. And again, they they have a massive advantage if they do come in that they are obviously well, they’re in everyone’s phone. Everyone has Revolut.

Well, not everyone but a significant percentage. So I would imagine that they’ll be a technology based version of where but they’re still going to need the advice piece. I still don’t know where and how they figure that out.

But again, more competition. But as it stands, the three main banks here do have, it’s not a monopoly, but it’s a duopoly.

22:40 – 22:44 Norman Schiever

What’s the triopoly? I have no idea.

22:44 – 24:08 John Coleman

A triumvirate, a triumvirate of probably about 80% of the markets, right? So it’s, it’s so getting them to move from, they all regard their market share very, very carefully, right? Now, unless the new banks that are coming in that have all these AI savings can then ultimately compete on a rate point of view, right?

When that day happens, when the slickest operational machine comes in, and is also offering the best rate in the market, then the big boys will have to move in line with that or else they lose market share. Yeah, very, very quick way. So for the moment, that’s not happening, because the slick banks are at the minute, are more expensive from ultimately, that price will, will win out in 90% of the case, if two banks are offering the same amount of money, price wins, if one bank is offering more money as a first time buyer, well, they’ll probably go with the more money if their need is immediate to buy a house, right?

Yeah. So, but if in time that they can become more competitive on a rate point view, then it is of only benefit to everybody, right? So yeah, but we’ll see where it all, we’ll see.

24:08 – 24:19 Norman Schiever

Yeah, well, I’m sure we’ll be doing one of these about AI again, and maybe some announcements from you and what you’re doing, maybe industry updates a lot sooner than one year from now, huh?

24:19 – 25:31 John Coleman

I would actually believe that will be because this world is changing at a rate of knots. And I definitely think we’ll be adding more to the, to the AI, what we would call it, someone’s used the term, a building a bot army. And the JC bot army, I don’t know, it’s, it’s certainly, it’s exciting, but I am, I’m still nervous about it in terms of, as I said, where does it all end, right?

And my view on it is, I want to help all the people that work with me. I want to make sure that they don’t feel threatened by it, okay? Because a lot of the things we’re doing is probably reducing the time that they would need. And then suddenly they’ll think, oh, there’ll be a need for me. Well, my view on this is, oh, absolutely there’ll be a need for you. Because hopefully the business plan here would be that we can do two, three times the level of business because of the systems, but the same level of customer support will be there with the team because they’re available, their time is available, you know? So yeah, that is the big picture plan, you know?

25:31 – 25:38 Norman Schiever

What’s fascinating? I love it. Yeah. The army, the robot army is coming, but we’re going to be friends with them. So we’ll be on the right side of it, huh?

25:41 – 25:59 John Coleman

Yeah, no, you sometimes, I don’t know if you’ve ever had this, when you might be researching a subject, right? And you’re already using, it’s also, I have to sometimes question my own sanity, right? I do sometimes feel I’m talking to a real person. I’m like, John, come on, this isn’t real.

25:59 – 26:33 Norman Schiever

Here’s something you laugh at, people, and I do this, it’s idiotic, but I’m also very old fashioned, where I say please and thank you. And they recently came out and I think it was ChatGPT or whoever the, you know, the head guy is, says it costs them tens of millions of dollars every month just to have built in extra responses and people having to filter through the please and thank you. And it’s like, they actually don’t want people to say please and thank you because it takes them longer and costs them more. But people like me, I can’t lose the politeness.

26:34 – 26:54 John Coleman

No, I’d be the same. And it’s funny, I even think it’s got my sense of humor. So I sometimes would have a joke with myself and put it in writing and make myself laugh. But it’s generating a response in a kind of a, okay, this guy changed my humor, which is even more scary.

 

26:54 – 27:10 Norman Schiever

That might be a question. Yeah, you might question your very robotic, machine like humor more than anything, huh? Well, this was awesome, John, but thank you. And Tea & Mortgage 29, anything else before we sign off?

27:11 – 27:37 John Coleman

From my end, not really, other than any questions, as always, looking for the roadmaps, obviously are viable. So if anyone wants a roadmap, my details will be with this. So just get in touch with any specific queries.

And we’ll take it from there. But I’m hoping that the world doesn’t change so fast that we need to do this again in three months. So every six months, I think it is a fair.

27:37 – 27:45 Norman Schiever

Six months is good. Yeah, we’ll be good to go. But all right, everyone, Tea & Mortgage, take care. See you next month. And then we’ll talk about this in six months.

27:46 – 27:48 John Coleman

Cheers, John. Thank you.