DayCast's discussion with Ron Frissora on how MI Homes is using Artificial Intelligence to accelerate productivity at MI Homes.
Announcer: Welcome to DayCast, where we discuss how AI is really used in business, no hype, discussions with Real AI Users. Here's your host Corey Dayhuff.
Corey: Welcome to DayCast, where we explore how AI is really being used in business. I’d like to welcome to DayCast Ron Frissora. Ron has been the CIO for M/I Homes for the past twenty-six years and has helped the company grow from a regional builder to a publicly traded builder, has built more than a hundred twenty thousand homes in fifteen markets throughout the United States, and M/I Homes is currently 12th largest builder in the United States. Welcome Ron.
Ron: Hey Corey, good morning how you doing
Corey: Not too bad for a chilly Columbus morning, almost in the winter here, which kind of stinks.
Ron: it does, but we got a beautiful day supposed to hit fifty-seven degrees today.
Corey: So almost golf weather.
Ron: think I might sneak out today. I doubt I'll be able to get a hole-in-one like you did, but I'm going to give a shot.
Corey: That's your goal right! Yeah, every time you step up there.
Ron: I'm not quite as good as you are.
Corey: Well good is a relative term. Yes, it is. So, you have kind of an interesting journey becoming a CIO. Can you tell us a little bit about that?
Ron: So, I grew up in New York and went to school and was a math and economics major decided to get my MBA right after school and got out of MBA School in the middle of a recession, and so it was pretty difficult to find a job. So, I was working for my dad and a diner on my feet, fourteen or fifteen hours a day hating every minute of it, and my brother called me. He worked for IBM and said I have a PC technician job open, there's no benefits, but you know we can get you in, and I said I have no idea what a PC technician does, but I'm interested sign me up. So, that started me into the world of technology by accident and I was able to leverage the skills that I learned through IBM and a couple other organizations to progress throughout my career, and I've been working for MI, as you indicated earlier for twenty-six years, started there as a business analyst for our mortgage company, putting in a new system for the mortgage company and gradually just work my way up and into the CIO role, and I've been in the CIO role for a little more than twenty years now.
Corey: So, you’re like the only other person I know that has had the same job for that length of time. I mean seems like in today's age max two years, and you switch jobs.
Ron: And as a CIO, right?
Corey: We must be really good at it.
Ron: I know CIO is supposed to mean career is over right, but so far, I've been lucky enough, not to prove that to be correct.
Corey: You're doing something right, obviously. Tell us a little bit about some of the projects going on at M/I homes around artificial intelligence.
Ron: I would say we're doing AI in really three different areas. I would say in robotic process automation, business intelligence and data loss prevention are really the three primary things that were working on today,
Corey: We'll start with RPA.
Ron: So, in the RPA space we are using UI path as a tool that we selected. To be honest with you, we started this as a skunkworks project because I had brought some of this to our steering committee and did not get any legs, worked with a couple of vendors and did some proof of concepts for relatively low cost, got our chief accounting officer interested in the product. She saw some opportunity within the accounting operations, where we could streamline a whole bunch of things, manual work that they were doing, and the two of us decided that we could leverage some funds that were available to do this behind the scenes and prove that this is a really good technology for our organization. Once we were able to do that, we brought that to the steering committee. We showed them the proof of concept, and it was approved very quickly. So today we have built many processes through RPA, where we're automating, redundant tasks that our employees are performing or were performing, and it's made a huge impact on our company already
Corey: Excellent. So, can you give me example of one of those?
Ron: Yes, so we own a mortgage and title operation as well, and so every time we close a home, the title operations have to report all the financial information that goes into our JD Edwards ERP system, and there's probably ten or twelve pieces of information that have to be keyed into the JD Edwards system to realize the revenue for that sale. In the past, that was a piece of paper that would be given to the accounting department and ultimately keyed into JD Edwards. We've leveraged AI, for now taking that information electronically out of our title system, leveraging OCR to grab the information, put it into UI Path and ultimately enter that information into JD Edwards, and it's been a huge success. This year will close you know, somewhere in the neighborhood of eight thousand homes. That task takes roughly ten minutes per entry into JD Edwards, so you can see the cost of the time savings that we've realized from that.
Corey: That’s outstanding.
Ron: And the other thing I point out is that the work that someone was doing was mundane and not really rewarding, and now they don't have to do it anymore. We built air checking into it. So, if something goes wrong, it tells us that something went wrong and that may be escalated to that person, who then can take care of that problem, but they've just got back a ton of time to be able to work on more meaningful things. And the people that were doing the work were highly paid people, so we've realized a significant benefit just in that one particular project,
Corey: And so, have you also seen a reduction in error rates since it was manually keyed before?
Ron: Yes, we have. We did have a fair amount of error checking in the manual process. But you know again when you go back to time savings, we are very accurate with what we're doing, and we don't have to do the manual work to check things anymore. So, I would say we've realized a significant benefit and accuracy and really improvement in our overall process.
Corey: So, the front part of that technology you talked about was OCR. OCR has been around for twenty years. No one thinks of that as an AI technology, but it really is. With the optical character recognition, you're really pairing an old school AI technology with a new AI BOT technology and now increasing the benefit you would have received off of just the OCR component of that to begin with.
Ron: That's completely correct and it's proven to be, you know very valuable to the organization.
Corey: Any other RPA success stories there that you're able to realize? I know it is kind of an ongoing process. Now you're identifying more and more right?
Ron: Yes, we have identified somewhere in the neighborhood of forty or fifty more processes that we know we can automate and, quite honestly, we've focused primarily in the corporate space so far. We have not spent a significant amount of time in the homebuilding operations, and that's kind of coming next. So, we formed an RPA steering committee, and that's how we're going to do some of the prioritization of the work that we're going to perform and really learn where there are opportunities within our operations to expand the use of RPA.
Corey: So, you went from hidden skunkworks project. You are kind of, you know, you knew it was good for the company, but you really couldn't get anybody in the company to pay much attention to it, to now they're, convinced and they're even willing to get on steering committees, because you have so much work now to do with it. When you talk to other CIOS, is that kind of a common theme that when talking about AI technology, it really starts as a small skunkworks?
Ron: I think more and more, especially with something new like AI, it's very difficult to go to a steering committee and say hey, I want to do this AI project. They've heard the buzz word they don't know what it means. They don't know how it can impact their business, and so we have to find ways to prove that it can improve our business, and so we have to show it to them. It's kind of like show me what you can do, and so that's I do think that a lot of things start that way, and a lot of I know that in talking to other CIOS, some of them have funds where they can just go out and play, and I would say in organization we don't necessarily have that. I mean I have some authority to do some things within my budget, but you know when I know things are right, the right thing, for our organization, and when my team knows that things are right for the organization we tend to go out and do them, and then we show what they can do and ultimately try to get the business to get interested in them. And that's really, I mean that exactly how we did this, and it proved out to be very successful.
Corey: You also dipped your toe in some of the business intelligence, AI predictive analytics stuff too?
Ron: Yes, that was a similar project and came about the same way. I’d been talking about BI for five plus years, I would say, and really was never able to get legs under that project. So, again I decided to go out and do it on our own, put a couple people on my team on this project and try to look at how we can present information to the company. Our three core metrics are starts, closings and sales, and what we did was we kind of just built some dashboards around that to be able to show that at any time you can just look at this information to determine the health of your company within seconds, and so that was kind of our entry into it. Once we got that going, then we, you know we were able to bring it to the steering committee, get the funding approved to advance the cause and go after more stuff. So today I would say in the in the AI space the thing that we are working on and I would say we have- we are not there yet. We are absolutely working on becoming more predictive in the use of BI and the biggest thing that we're working on right now is: How do we leverage BI to predict future sales? We know that there is a correlation between the number of leads we get from our website, the number of traffic units that come to our model homes and, ultimately, how many sales occur within the next thirty to sixty days. We know there's a correlation. We don't know exactly what the correlation is and we're trying to learn what that is and then be able to predict what our future sales will be. I'm sure that there's going to be other things, for example, interest rates- or you know, unemployment rate or things like that that play a role in our future sales. But what we want to do is start small show that there is a correlation, start to predict what our future sales are, measure the accuracy and then factor in other things, to get more precise.
Corey: And have any of the tests been successful?
Ron: I would say it's a little early yet. We’re still in the working on stage at this point, but you know because of the journey with BI has been several years, but it took us a while to kind of get the interest of the business, and once we were able to show the value, then more things started coming in. So, now this is kind of the next evolution of our journey, and this is hopefully going to be the one that proves to be you know the most meaningful, I think.
Corey: I think we've seen a lot of companies kind of follow the same path that the you have when they start with RPA, because there's usually no brainer cases within every organization, where you can leverage robotic process automation and automate some of those remaining mundane tasks that are out there and then once they get a taste of what AI can do, the next logical thing is the predictive analytics around you know, whatever BI system that they currently have. It just seems to be kind of a trend that almost every organization goes through and then, after that, it kind of takes on a life of its own, where you have multiple things going on in multiple areas within IT that are all related to AI, all not the same, because AI means a lot of different things to a lot of different areas.
Ron: Sure, that's absolutely what we've seen. And I will tell you that now we're struggling to keep up with the demand and so we're definitely going to be looking to at more people in this space, because we see a tremendous amount of value, we're getting a lot of demand from the business, and we need to produce continue to produce great results.
Corey: Sure, now comes the tough part, producing.
Ron: Yes, exactly
Corey: So, any other projects, in an any other areas that that you're looking at AI, or currently using AI?
Ron: You know, I would say Data Lost Prevention is another area. We own mortgage and title operations. We are a public company. You know, obviously, we have a lot of information within our four walls that is very confidential and we have to protect our customers’ data, and we take that very seriously. So, we've implemented some technologies that look at data lost prevention and determining, you know, are there things leaving our premises that should not be leaving our premises? And, interestingly, we put some tools in place to measure that, and we really just wanted to see did we have a problem? And sure enough we had some things that were being sent through email, for example, social security number or driver's license numbers and information like that, and although our policy is that you're not allowed to do that right, I think just some people don't really think about it. They don't understand a lot of customers send us that information through email, and you know despite the fact that we have secure tools to be able to have them just post something versus sending it to us, via email. So, you know we did study it, and found that there were some things going out that you know we needed to address and you know, then we put better policies in place around that and then also put prevention in there, so that things could not leave. And it's proven to be you know very successful and we've been able to educate our employees as to why you don't do these things, and it’s been a big impact.
Corey: So funny story. Yesterday I was on Facebook and, like my great aunt, obviously got a new cell phone. She posted got a new cell phone. Here's my cell number to the entire world. If you didn't have spam before you are definitely getting spam calls now.
Ron: Yes, you are. People don't realize that information just propagates so fast and you do get spammed or you try to get scammed.
Corey: Yeah, I've had the same cell number for twenty years. You know it's out there on I don't know how many lists but I use as many of those spam blockers as I can possibly find. I still get you know fifteen calls a day on getting new auto insurance or updating my warranty. So, in your AI journey like where did the users really start to get on board and see the benefits?
Ron: I would say when they see the value. So, good example on the in the RPA space, most people when I show them a demo of this, they don't immediately understand. They can see the value, they can see that it's going to take people less time, but they struggle to think about how does that apply to me? But once they go back to their business and they see all of these mundane repeatable tasks that people are performing every single day of the week, they then come back and go hey that RPA thing, do you think it could automate this for me? And generally, when we prove that it can, you know, then they start thinking about I'm going to save all of this time for this person. How am I going to reshape that person's job, to do more, to you know to impact the business more rather than keying data into a system? And it's proven to be incredibly valuable to the operations.
Corey: You said you kind of brought some RPA companies in, you ran some prototypes ahead of time. I know from our experience in going into talking organizations, especially around AI, probably the most difficult part is explaining what AI is and can do, right? Because the chasm between the marketing buzz words and reality of how it gets implemented within a specific organization for specific activities in the ROI, it's really hard to describe for the user what that experience will look like in the end, and the value of it. I don't know, it's probably less about them not believing that it can be done and more about they just literally can't see the vision of how that happens.
Ron: Exactly right, and I will tell you I don't use the buzz words until the end. Then I tell them you know this is what we're using to accomplish this. But I ask them about their problems. You know, tell me where you're spending a lot of time keying data into a system, you know, and it's a business problem. They may not think it's a business problem because that's just what they do right? They have to do this task in order to satisfy our accounting needs, for example, or HR needs, and they just they have to perform these things. So, it's just part of their daily business. So, I don't ask those questions you know, or I don't share with them information about AI or RPA. I ask like tell me about mundane tasks that you're doing in your operation, that take you a lot of time. And then I show them a video that we put together that shows, hey guess what, I can have a robot do that for you. And guess what, it'll be really accurate, it'll be logged into the system, just like an end user, and I can audit it. So, if anything ever goes wrong, I can go back and tell you what happened and why it went wrong. And I can give you error checking. So, if something goes wrong, I could leverage the person that was doing it previously, send them an email and say you need to look at something here and go fix it. That's when they, the light, goes off and they see there's incredible value to do this.
Corey: What always amazed me when going in and talking to customers over the twenty-three years I've been doing IT no matter what technology it was, AI or whatever it was, if it was a new technology that they really hadn't dealt with yet and trying to explain what it does, quickly, they jump to the detail on well how's it doing that? Sometimes I didn't even know. My standard answer would have been you know its automagic. It's a lot of coding underneath that makes it happen. I don't need to know all the ingredients that go into a good cake if it’s good.
Ron: It just works.
Corey: Right so, technology layered on technology layered on technology. I don't need to know how the CPU fits into the motherboard to understand that RPA is going to solve this business problem for you.
Ron: Yeah, and that's our job to figure that piece out right. You just tell me about your business problems and I'll solve them for you.
Corey: It's almost like they have that desire to know. I'm not gonna trust this thing until I know exactly how it works underneath which in almost any other business in there, they wouldn't ask that many questions, but it seems like when it comes to technology and AI specifically, they really have that interest.
Ron: And I think you know for MI homes as a publicly traded company, I think it's important for people to ask those questions right. We’re bound to Sarbanes-Oxley and Gramm-Leach-Bliley and all these regulations, and the reality is we have to understand those things. So, I think we have our arms around it, now, since we've been doing this for a while, and we're able to, you know, communicate that information to our workforce to satisfy them. I think showing them is really the biggest thing that gets them convinced that this is the right thing to do.
Corey: Sure. Would you have done anything differently in retrospect, in this journey that you've gone through?
Ron: I may have started in the business operations instead of accounting and the reason why I say that is when we can impact business operations, we get more interest. It doesn't mean that accounting is not important, and streamlining a lot of the things that they do, but generally when I can show that I can save time for the business operations or make them more money, people's ears perked up a little bit more.
Corey: Sure. From an IT organization perspective, did it change kind of operationally how your structured or how you function? Is there, after adding to your quiver of capabilities AI, how did it change IT?
Ron: So, we've restructured, we've added people and then we've restructured people within our group to work on these things. So, I'll give you an example in the BI space we started getting a ton of demand, and so we realized that we primarily were working without some outside consultants and we primarily had one person who was working in the BI space. We quickly realized that one person was not going to be able to keep up with the demand and if we, you know, continue to leverage outside resources, we were going to spend a lot of money on it. So, we ultimately ended up shifting a couple people within our group into the BI space to handle the demand. We now have a steering committee around business intelligence and with probably ten people throughout the organization, both leadership and people who are you know, controllers and things like that, that are studying a lot of the business and trying to make improvements in the business. We follow an Agile methodology there. So, every two weeks we are delivering something new in the BI space, and it may be small or it may be big right, but every two weeks something new comes out that we distribute to the business to make things a little bit better and just provide more insight. We did the same thing in the RPA space, so I would say we are definitely leveraging a little bit more outside support for that, but we shifted a couple of people from our help desk who expressed some interest in you know doing more development work and got them to get going into the RPA space. I will say one thing on the RPA. We can with low code no code type of stuff, it's very easy for someone with a technology aptitude to be able to jump in and pick up some of this work. The individual who happens to be leading the charge for us in the RPA space has an accounting background. She was an internal auditor for a while. She came over to our group as a business analyst, never did any coding in her life, and now she's very proficient in the tool.
Corey: That probably gives her a completely different perspective than if you just pluck somebody out of IT to do it.
Ron: There's no doubt. And the great thing about her and others in our group is she has the business background, so she's able to look at business process and determine how do we take that business process, possibly reshape it to then fit into an RPA tool.
Corey: Sure. The most interesting thing about where IT has gone over the last five years and really accelerating into the future, is pushing the development cycle closer to the end user, with the low code, no code capabilities, with the self-help analytics kind of capabilities and so forth, and you can go on and on. It’s all being pushed closer and closer to the business, which is a good thing.
Ron: It's a blessing and a curse.
Corey: That’s true. All the problems still come to you.
Ron: Yes, and then you tend to lose control of some things that you find out about later. Sometimes you learn some things that are going on in the business that you weren't aware of and then guess what it ends up on your plate.
Corey: Where do you think, compared to other builders, where do you think you guys are from a technology perspective? Are you guys leading the charge there? Do you think you are kind of in the middle?
Ron: I think we have the best technology operations in the home building industry. Just kidding. I think we do a really good job. You know I'll just say I have a great team and a lot of seniority in my team. We've had people that have been there for twenty plus years and they do a fabulous job, but I do participate in a homebuilder CIO meeting every year, and we were very open. We share a lot of information. It's a great group of people, and I can go to them at any time if I have, you know, issues that I'm dealing with that are very homebuilding specific and get a lot of support from that group. So, but you know we look at the things that we're doing as an organization, and I can tell you that, from the perspective of just technology operations and the I'm going to call it integration of data throughout our back office systems, I think MI is really at the top of that. When you talk about some of the newer technologies, I think that some of the other builders, some of the larger builders entered the technology space, perhaps a little bit faster than us. I would say we're a fast follower from that perspective and I think that we, you know we tend to kind of catch up, pretty quickly and really stick with the pack. So, I would tell you in the RPA space, we were one of the leaders in that space. We started earlier than just about any other builder. I think a couple of builders started dabbling into it. We've had separate calls specifically about RPA, where I know that at the time that we started, we were one of the only builders that were doing it, so I would say we were a leader in that particular space. In the BI space, I would say we were a little behind. It took me a little bit of time to get that approved through my company, but you know now that we've gotten it, we've done a fabulous job of really catching up. So, I'd say we lead in some spaces, we follow, in others, but overall, I'll put our technology up against anyone else's.
Corey: Excellent. I would too. Well, I want to thank you for coming in. I really appreciate you doing this. Some great conversation, some real good insight. You run a great organization, and I hope you get to enjoy your golf today.
Ron: Thanks Corey appreciate it. Good seeing you today.