The Mark Perlberg CPA Podcast

EP 036 - The Revolution of AI and Automation in Real Estate w/ Chris Tamm

Mark

Send us a text

Want to explore artificial intelligence and automation? This episode features AI expert Chris Tamm, who shares insights on AI's impact on real estate, including content creation and data extraction. We also discuss AI's natural language capabilities and robotic process automation, offering a glimpse into a future without mundane tasks.

Our journey explores data analysis possibilities, such as AI cleaning data, asking questions, and drawing inferences. Imagine chatbots simplifying search processes. We also cover robotic automation for tasks like data collection, potentially changing the game for rental property owners.

We conclude with thoughts on the future of robotics and AI, their revolutionary effects on work, and reflections on the potential transformations they can bring across industries. Get ready for an enlightening discussion on AI's exciting opportunities.

About Chris:
Chris Tamm has over 25 years of leadership experience in various roles, including real estate and insurance. As Founder/CEO of Ownership Mortgage and Ownership Insurance, he's helped many achieve financial goals. He's also the founder of Cast Services, an AI and automation firm. Learn more at www.christamm.com.

Mark Perlberg:

So welcome everybody. I'm here to be joined by Chris Tam, really excited to jump into a conversation on automation, artificial intelligence and how you can use this in your business and, in particular, how you can use this if you are in real estate. So, chris, why don't you give us an intro of who you are and what you're doing in 60 seconds or less, and we'll jump into some conversation?

Chris Tamm:

Yeah, mark, I appreciate you having me here, lesson 60 seconds. So I'm an entrepreneur by trade. I built a man one of the larger top 20 real estate teams in the nation, built one of the CRMs and residential real estate space Firepoint. I operationally run a CEO of Mortgage and an insurance company and they started at partner with real estate agents and we do AI and automation conversions for other companies. So install these types of tools in a lot of different verticals.

Mark Perlberg:

Wonderful, and so right now, ai is a super popular and sexy topic, and all business owners I know when entrepreneurs are looking into some of the basic stuff with chat, gpt and Jasper and all that stuff, and it's great for brainstorming and content writing. But what would you say? So I would say the masses are already picking up on using AI to do some research and maybe write some paragraphs, and some of the teenagers are probably using them to write their essays and their cook. I don't think academia will ever be the same again. So obviously we have all of that going on, but where are some things that you think people don't realize is available for them in the world of AI?

Chris Tamm:

That's a great question. As far as what they don't realize is available, depending on the skill level and what they wanna do, I think some pieces are what you can fairly easily connect to when it comes to like machine learning, teaching databases, how to learn and answer things back. There's lots of tools that have rolled out in the last six months or so that make that a lot easier, depending if they don't wanna go super technical, even ways where you can kind of drag and drop huge tons of information or point these chatbots at websites and have them learn all the content immediately so you can just talk with it. So that's starting to solve some challenges around people digging for information and going through these massive PDFs or Word documents, kind of finding answers. Another one is it's tricky to do but it's available which is extracting information from pictures and videos and so, like we talked about some of that, we can show some of that in a little bit.

Chris Tamm:

As far as pulling even like very specific details that have quality of things, types of houses, I mean, if we talk about real estate specifically, even like what types of countertops or quality or lighting levels, some other ones, video has lagged a little bit behind where we thought it would be by now. So like we don't see, unless it's super expensive, like the quarter million dollar your contracts. We don't see real-time video entering yet. But process video, where you can have a system create video from scratch, has never existed, is making really good strides, but it hasn't aired the mainstream yet. And then I think maybe another piece that a lot of people don't know about but has been around for quite a while is robotic process automation, which is computers using computers and so letting a computer do the tasks that a person normally would like navigating a computer using Excel, using traditional programs, and that's making pretty fast strides also.

Mark Perlberg:

Interesting. Tell me about. So tell me about the robotic process automation, that last one that you just saw. What did you say? It was called again.

Chris Tamm:

Yeah, RPA so robotic process automation.

Mark Perlberg:

Okay, so tell me about some real-world applications of this that you're seeing.

Chris Tamm:

Yeah, so some of them. I'm trying to see if I have a quick one that I could pull up for everybody here, but it's a couple of different ways. It's basically having a. It's a subset of AI, so it's one of the pieces under AI. It's computers that use other computers, and so there's some simple versions of this, like there's one let me see whether it browse AI. I think Otter's another one, but these are like traditionally screen scraping tools, and so if you wanna get a lot of information off the web, you can point these tools at something and say, hey, extract all this information from these pages, go to the next page, extract it all, throw an Excel for me or throw it into a Google Sheet Now, and so you start having computers, use computers, and then you can teach them to do things like you can teach them to use Excel or to use QuickBooks, like QuickBooks desktop or other programs, and so they can actually start doing tasks for you.

Chris Tamm:

There's ways to they probably goes a little too technical from an implementation standpoint, but there's ways to just type natural language into your computer. They call it. It's a program that was started called LangChain. It was an MIT program, but you can just type instructions to your computer and your computer will then run your computer and go download programs and find the ways to accomplish that using AI. So it actually starts taking over your computer, navigating websites, opening Excel and actually saving files and running your computer for you.

Mark Perlberg:

Hey, very cool. So now I'm thinking to myself let's say I wanted to, so do you think I can teach my computer to go online and copy and paste all the articles related to a certain tax code? So I say, hey, go to the web, copy and paste, pull out, put this all into a word dot, sorted by relevance, et cetera, et cetera.

Chris Tamm:

Yeah, so I don't from right where it's usable today. It's getting better with training and with specifically implementing. You can get it to do that right off the web If you had to download it right now and do that. It's not perfect yet. That's kind of where it's the bleeding edge, where it's getting better. But if you knew how to use some of the tools like there's like automate anywhere UI path, there's some tools where you can do a little bit more integration or deeper programming with it, yeah, so you could definitely get it to do that. Like part of what our team does when businesses hire us is we kind of fine tune and implement whatever specific tool they need to do that kind of en masse. It's not. It's not there yet to where anybody can just download a type and have it run everything, but it's getting close to that.

Mark Perlberg:

Interesting, interesting. So in the world of real estate investing and being a real estate agent, we'll talk about the tool you've created a little bit, but where are some ways that folks that listen to our podcast and web of our series, where are some tasks that you think that they could probably automate if they have a deeper understanding of all the resources available for AI?

Chris Tamm:

Yeah, as far as today, and that's been kind of like the one of those really big questions is outside of chat GPT, you're outside of creating an image on mid-journey. Where can you use this right now? Like to save time? I think some of those. So manipulating Excel is one of the areas that a lot of people don't use chat GPT for. But even manipulating Excel, so using Excel to do forecasts, to do analysis or to do goal seeking Some of those are as simple as like copy and paste Excel information directly into chat GPT and ask it hey, can you forecast this for me for the next six months?

Chris Tamm:

Or if my costs go up by 10%, my income goes down by 10% and I lose a renter for a month. I can't project this out for the next two years, and it'll do all of that almost instantly. You can also have it goal seek the other way. So here's all the information I have. What do I need to do? Give me five options of how to make a million dollars with this in the next 18 months and it'll give you all the variance analysis. Give you all the options. So doing analysis, doing forecasting, is one Using it to go forward.

Mark Perlberg:

Chat GPT interface with Excel, though if they wanted to do some projections, we'd also create charts and stuff that would rely on using Excel to put together, to use all the data and create visuals. Would chat GPT be able to do that?

Chris Tamm:

So in a sense there is a way to do it. There's a couple programs that bring AI into Excel. There's some Excel plugins for it. I'm trying to think the easiest way to deal with it. One of them is going to be Microsoft. Microsoft is rolling out their Microsoft co-pilot, which is going to be in all of their applications, and that's not fully live yet, but in the very near future they're going to have AI integrated directly into the applications and potentially even cross application, so you can just have it do things directly. There are a couple of plugins that connect Chat GPT to Excel, and then there are a couple that connect it directly to Google Sheets so it can access those.

Chris Tamm:

I would say they're not 100% bulletproof right now. Another way to deal with it is and this is a little bit of a roundabout, but I think it's an interesting one is you can go into Chat GPT and tell it what you want to do with your data. You can tell it about your data in Excel, like hey, I've got a list of 10,000 prospects or 10,000 houses or properties I'm analyzing, or a property list or a tax list, and you can tell it what you want to do with that. Like hey, customers in column one statuses in column D, and I want you to sort all of them, remove anybody with something in column E and then put them into a table and format them a certain way. If you tell it about your data, then Chat GPT can kick out the visual basic and tell you how to run that in Excel, and so it's one extra step, but it builds the code for you and then it tells you how to run it and that takes over your Excel and builds everything that you want it to.

Mark Perlberg:

Now here's another thing I'm thinking about here and we're going to probably do a separate meeting with all of our clients on this and we get a ton of data to try and make sense of all the data coming through, all of the systems that are interfacing with QuickBooks Online.

Mark Perlberg:

And then non-financial data, and we have tons of short-term rail investors getting tons and tons of data from Airbnb, the RBO and all sorts of third-party vendors, and if we could find a way to aggregate this data, to capture different patterns, to assess profitability and to understand maybe some adjustments that need to be made, I mean, there's just tons and tons and tons of data and I imagine there's a way that we can leverage AI to make correlations between different items as seen unrelated, or they are related. Maybe we just can't make those connections yet, and so some of the things I think about for our clients are what is the? You know what's the average lead to stay and how does that affect profitability and booking rate, vacancies, rates, and how did the cleaning costs compared to the cleaning fees, based on different types of engagements or all sorts of patterns in correlations, being connected? Maybe we wouldn't be able to recognize with our naked eyes on this data.

Chris Tamm:

Yeah, very much. One of the things that we see is it's getting cheaper, faster, easier to build, kind of what you said, the idea of like a data mesh or data fabric. When a system can see all those different data points, it just looks at it as one piece of fabric and then, yeah, you can either ask it to draw out inferences or correlations that it sees, that you can kind of take a look at and see what they mean. And then there's also quite a few different like business information or BI tools that you can plug it into as well, whether it's something like off the shelf, like a Microsoft Power BI which has an AI natural language. You can just ask it questions about the data and understand what's there. But also, like you said, like the benchmarking piece, like if you can upload a tremendous amount of data and then start having it come up with inferences and bring in new data and learn about it. Yeah, a lot of ways to use data that way.

Mark Perlberg:

And one of the challenges in trying to access the data is just cleaning up the data and formatting the data and finding formulas that will maybe remove the extra spaces and all the extra things. So, instead of racking your brain on all what kind of formulas going to do exactly this, this and this, you can probably give it to AI and they'll run a couple hundred scenarios to eventually arrive at the perfect formula to give you the data in the way that you'd like to look at it as well.

Chris Tamm:

Yeah, and that can be done even with some people, just directly through chat Gbt, like if you know what's wrong with your data or you know what you need to get rid of it. There are ways to, yeah, even go into chat Gbt and say, hey, I've got an Excel file or CSV file nine, to get rid of all the commas and these, all the characters that are non numbers in this column, and you just tell it what you want it to do and then it can spit out the code that it'll just take care of and do all of that for you in Excel. Again to your point, getting those ones that are integrated directly with Excel or directly with Google Sheets, where I can just do it for you directly, there is easier, but there's ways to add one step and just do it through like chat Gbt right now.

Mark Perlberg:

So one thing I and one thing I noticed was Google has their new AI coming in that's going to help you, you know, respond to emails and will likely and will interface with Google Sheets. So that's another. I don't know exactly when it'll be available, or if it already is or tested out, but you know, all this stuff is really close to being in our fingertips.

Chris Tamm:

Yeah, and there's other is. There's some other ones like a I think it's personal dot AI, where it'll build kind of an AI machine learning around you to where you could plug that into your text message in your email and have it right now auto suggest what you should respond with, and you can kind of create different personalities you can have a business personality or personal one and actually upload different information to it.

Mark Perlberg:

So what I'm wondering now also is when you're let's say, you're trying to research some recent events, you know, with chat, gbt is not up to date because I think it's like it only has data as much as going back like two years ago at least, as of right now, right? So if you want to research a textbook or all you know, the textbook would be a change or any other current events that you're looking to gather information on. So if you're looking to ask a question and get a direct answer that sources other websites, I think you can now do it from Bing, because Bing is tied directly to their search engine. But what are your thoughts? If we want to get accurate, up to date, we want to get an answer that's drawing from current events and the most up to date collective wisdom of the Internet.

Chris Tamm:

That's a great question and again, these are all I would say, well educated opinions from our entire network and what we do. A couple of different answers to that one. Chat Gbt has the Bing search with Bing integration. They pulled that back over the last, I think, week or so and we're waiting to see when it gets re-released. And so one way, once that's re-released, is to go into chat Gbt, select the GPT for and then select search with Bing, because then it uses the chat GPT large language models and GPT process, but then it also uses Bing, so it's actually clicking on links and searching and going through and pulling information together along with chat Gbt to answer. So that will take who knows when they re-release that piece.

Chris Tamm:

One other option is to go in and use specific plugins inside of chat Gbt. So, depending on what you're looking for, like there's a plugin what is this one I've used it before which is scrolling through this like scholar, ai and Wolfram, where those plugins have live access to their programs in the Internet. And so if you're looking for like scholarly articles or specific things, you can use the plugins in chat Gbt and those plugins do have real time knowledge. If not that, then I would recommend kind of what you said of using Google, using Bing, where they are using AI to pull in real information and to do some of the natural language pieces, and you can do a lot more on that now through Bing and Google than you could, I mean, obviously, a year ago.

Mark Perlberg:

Wonderful, very cool. So so it looks like you know, we have like a personal assistant that'll give us, that can now give us references, and I remember interfacing with chat Gbt, you know, like, can you tell us where we got this information? And they would just source me to links that were out of date. But now with the big chat, we can actually get source references for those items, which is pretty neat, and so we can do the research, we can pull data. I'm you know, sometimes I think that you know there are some manual tasks that is still going to be grown work. You still got to either do it yourself or hand it off, and before me is talking about your tool.

Mark Perlberg:

I'm wondering for folks that are investing into real estate and doing this stuff. I mean, they can use some of the other things I'm thinking about here use AI to write the non-disclosure agreements, maybe help with reviewing contracts. I know OCRs has its limits. You know we'll talk about how AI can help you out with creating articles, in evaluating photos, writing a newsletter you know, seasonal newsletters and all that good stuff doing reports on. Perhaps you can also even use AI to do reports on the market condition and trends.

Chris Tamm:

Yeah, and we've actually we've seen some people that like there are some open sources that will provide, like, basic market info. But there's ways through and again, you can do it manually. But you can also connect some things through like a low code or a no code environment, like a zappiercom, where you could take, like, when an email comes in or when something comes in to say a Google sheet or somebody inquires about something, you could automatically have that come in, pull that market info and then you can drop market info, somebody's LinkedIn profile information about what they said, and then you can have something like chatGVT, then merge all that into a coherent email and then send that email automatically, and so or send it for review so that somebody can read it first before it gets sent, and so there are ways to take some of these and combine it to where the steps that you need to do are more review steps as human in the middle, instead of somebody having to actually manually go pull all that information together.

Mark Perlberg:

Right. Well, one of the things I'm not thinking now. This may be too complex I imagine there's already a tool that does this but our clients, especially because we have so many short-term rental investors there's a lot of thought behind the strategy and pricing and the offerings and these properties and the pricing strategies depend on what's the vacancies in the neighborhood, what's the seasonality and how long are people going to stay at different rates. There's hundreds and hundreds of variables in determining how are you going to price this, are you going to offer multi-day discounts, and what's going to give you the optimal amount of profitability? And I imagine to use AI to help refine and optimize your pricing strategy is going to take a decent amount of research and tweaking and back and forth with the system. So what are your thoughts on how AI could be implemented in assisting maybe some of our clients in their pricing strategies?

Chris Tamm:

Yeah, I mean I think definitely can and definitely does in some aspects. I know some people that have worked on projects specifically in those areas. I can't say exactly where they are with rolling them out or using them internally, but it definitely can. I mean the idea of pulling information together, even from having these kind of robots, like the robotic process automation, go out and scrape other sites and specifically look for other information and pull reports, do analysis and make recommendations. It takes some tweaking and some training to get that right, so you kind of run that with human side by side for a little bit, but it's very doable. I mean a lot of these. If it's less than 10,000 rules that we would run in our own head, a computer can absolutely do it. It's more teaching it and getting that process right. But yeah, I mean almost anything that you can think of can be done now, just putting the steps in the right place in the right order.

Mark Perlberg:

Do you think it would be possible for, let's say, someone who is less sophisticated here, maybe has 10 rental properties and isn't quite ready to engage with, let's say, this is a DIY type of person do you think that they could potentially find a tool that's going to scrape all the prices and out on the Airbnb and the Airbnb websites and deliver them a report of the vacancy rates and what people are charging in their neighborhood at a certain point in the week? Do you think that would be possible with the current availability of tools we have? I think so. It's going to take a.

Chris Tamm:

Okay, okay, okay, okay, okay, okay, okay, okay, okay, okay, okay, okay, okay, okay. It'll take a little bit of time to maybe put it together and have it create the right report for them. If they're not very techy, there are, like even just doing that as a search, like scrape all Airbnb prices and vacancy rates. There's like Appify, which is an Airbnb data scraper. There's Web Automation, which has an Airbnb date, availability and price scraper, and so there are systems that are built right now to specifically pull all that off. But it probably would take a couple minutes of time to give those couple different tools and then be able to have that information come in so you can see it the right way. But with a little bit of setup you could clean that process up also.

Mark Perlberg:

Yeah, you could probably pull some data, make some hits, you know, pull some historical trends to generate and then maybe create some interfaces between that and how you're going to implement your price, you know, communicate or directly enter your pricing rules into the system, or maybe even and if you can't do that, at least create something that you can that'll show the instructions for you or your assistant to do so. I imagine, with you know, there's so much creativity involved in finding solutions as well.

Chris Tamm:

It's a very right brain type of activity, yeah it is and, to your point, how fast it changes. Like if we had this conversation six months ago, none of those tools may have been available. There's so many things that are popping up now because it's cheaper, faster, easier to put them together that even other developers putting things together like specifically we just asked for. For all we know, that tool is out here already. It may just not have shown up in the first five things I saw on Google, but the speed that these new tools that are rolling out is dizzying right now. I mean we see our team sees anywhere from well. I personally get funneled from our team up to 20 a day of tools that are changing or that I need to learn about, and there's, I mean, 20 to 100 a day that are actually rolling out or changing.

Mark Perlberg:

By any chance? Do you know of what the IRS is going to be doing to spot red flags and to go after some of the people who are business owners and entrepreneurs and real estate investors filing the taxes?

Chris Tamm:

I wish I had a connection at the IRS to have those types of answers, but I'm pretty sure if I did I probably couldn't say I'm here but no, I have no idea. Yeah, that'd be great.

Mark Perlberg:

I'm pretty sure I've heard that they're going to be doing something and it would be interesting to see what they're looking out for, where they're, where they're flagging and how our actions impact their algorithms. Yeah, that's an interesting one.

Chris Tamm:

I mean side note, not directly related to that, but there's a lot of new stories. Over the last six months, the UK has rolled out a project called the sovereign AI project and again doesn't relate obviously, to us here, but it's an idea of, maybe a portence of, what's to come. But there the UK is building an AI system that basically understands and can run their entire government. The idea that it could understand and see every detail about every citizen and every business right away and make its own inferences, recommendations, take care of its own paperwork, send out its own notifications. You can still have humans in the middle wherever you want to but in a sense something like that could do everything all at the same time, all at once, instead of this process now where we have to kind of I mean, there's a lot of manual process in the current system- yeah, where are you most excited about AI emerging technologies?

Chris Tamm:

Great question.

Chris Tamm:

I think it's probably two things. One is robotics, which is kind of one of the next waves title waves that's going to come on the shore, from how AI is making robotics so much more attainable and easier to program and do what we want with is going to, I think, change how all of us think about robotics and home robotics. And then the other one is more of kind of this process is, as we go through this true kind of revolution, how we think about, how we do our work, how we accomplish our tasks. It's what we come up with, like the questions you just asked, like is there a tool there?

Chris Tamm:

If it's not, somebody's going to create it see, they can be you or me or somebody else. So it's the conversation that I think that's what actually excites me the most. Is the conversation, the possibilities, the number of things that we've heard for the first time that didn't exist and then either we or they or somebody was able to build it. There's a lot of first times right now and those are those light bulb moments, because once I mean the thing that you want to build, I want to build off of that, because I don't even know what that is yet.

Mark Perlberg:

So can you elaborate on what home robotics and robotics is? Because, to my understanding, ai is the production of answers and knowledge and just tapping into these resources, but the robotics is the stuff that actually performs actions. It uses the data and actually enters it and creates different actions, but I don't really think I have the best definition. Is you tell me a little more about what robotics and robotics is and what it potentially do?

Chris Tamm:

Yeah, and I think a piece of that is a lot of people right now, and I think of the word AI, are focused on this text interface because of, like, the emotional win of chat GPT. So really, from an AI standpoint, there's all these little pieces that are inside of it and, like that, that chat GPT is a large language model and it's a generative pre train transformer. It's a demonstration of those two things. And I think for robotics, what really excites me is when you think of robots, the actual motors and the actual plastic or metal aren't necessarily that expensive. What's really expensive and what takes a long time is the computer system of programming a robot, and that's been out of reach for most of us for a very long time.

Chris Tamm:

Princeton had a recent example where they used chat GPT to program a robot and so they could just type to this robot hey, can you go pick up everything on my floor that's dirty and put it in that bin and then go grab me a coke out of the fridge.

Chris Tamm:

And this robot that was not trained by a Boston Dynamics, there was no millions of dollars that went in.

Chris Tamm:

It was a bunch of motors and sensors attached. They could program itself to go do those actions and so, as that starts to roll out in a home environment if I can go ask something, to go lay that cord on my wall or go paint that wall or grab me a coke and then go pull the weeds out of my yard and then go take care of my laundry the demand for robots is greater than one per person, and so it's one of the largest demands we've ever thought of from an actual product that we sell and the robots themselves that you think of Tesla. Tesla is really an AI software company. I mean, that's what their value is on this control system, because if you had that, you could plug that into any car or any robot, in a sense, and so just very excited to see robots roll out and the programming roll out to where it's as easy as chat, gpt or even just tell it what we want it to do and, whether it's done it before or not, it can learn and figure out how to go do it.

Mark Perlberg:

Kind of like if Siri and Alexa were able to move around and do things and a little bit smarter.

Chris Tamm:

Yeah, and the fact that another piece of there which freaks some people out but to me it's actually much safer is we used to think that all these things have to be connected to the internet, but they don't. The way these AI chips have rolled out, you can now have these chips that have their own machine learning, their own machine vision, and they're local and it's like a $5, $10 microchip, and so the idea that a robot could be directed by you and learn how to do things. It would never have to communicate anywhere else, it would never be connected to the internet. It's never talking to an outside system, so you don't have to be afraid of somebody else being inside your house having control of. These. Could be offline systems that actually still learn themselves.

Mark Perlberg:

Hmm, interesting, when we have all of these automations and things of place and I'm thinking about, I mean, the future is people are gonna be telling Alexa to get them a beer, and that's gonna be really fascinating to see the next revolution. And I think about the analogy of what Ella Beer's first came available and the buttons came. I'm not sure you ever heard about this, but the initial Ella Beer's there was someone who would come and operate the Ella Beer. That was his or her job to operate the Ella Beer. But when they first introduced the buttons to do it themselves, people were like I'm not touching this Ella Beer.

Mark Perlberg:

This is weird. There's a, you just press a button and there's no one to operate it and open the door. I can't trust this. And now it's like common day. So it's interesting to see when we introduce robots eventually to our home to clean our and we already have robotic back and we're gonna be able to do more of these things. It's gonna be interesting to see the resistance and how weird and uncomfortable it is after us as we get used to all these ideas?

Chris Tamm:

No, very much agree. I mean the societal change that we're starting to go through now. I mean a lot of these tools we've been able to use for two, five, 10 years Like we've seen those robots that can cook you dinner have been around for three, four, five years now. Obviously they're super expensive, so they're not mass adoption yet, but everything's starting to get less and less expensive.

Chris Tamm:

But yeah, I mean the societal adjustments we're going through is it was weird for me at a hotel what a couple of months ago to see kind of a symphony of these robotic lawn mowers mowing the lawn at the hotel. I mean just seeing all the things start to change. But yeah, we see it from like a business standpoint, as business owners is there as we're implementing for them these new AI and automation tools. There's a cultural change there as far as people readjusting to how to do a job, even team members adjusting to hey, I don't have to do that 80% of stuff I hated before, but do I want to do that 20% thing five times as much? I mean there's lots of adjustments at home. Personally, how we think about it, it's going to be a very interesting five, 10, 20 years.

Mark Perlberg:

So here's another thing I'm thinking about as a CPA. So if you were building tools and you were developing, and even if you're not spending actual costs on materials and supplies but you're hiring staff to do stuff and the payroll costs to develop and research and build these tools and create automations for your business and develop new ideas like this to potentially qualify for the research and development tax credit. And if you look deeper into the R&D tax credit, I mean there are companies making millions and billions of dollars in R&D tax credits. From what I hear is that Tesla.

Mark Perlberg:

Even if they operate at a loss, one of their greatest economic benefits of existing is just the R&D tax credits alone and sometimes you get some tax credits. There's a whole other world and this is outside of my super-vexed teens but there's a whole world out there of strategizing and maximizing your R&D tax credits in your business as business owners. So lots of really interesting and exciting things there and some of these tax credits explain part of why the super wealthy in order of these gigantic companies are not paying any taxes at all because of the tax incentives to use technology and further develop your business and your processes.

Chris Tamm:

That's a really good point.

Chris Tamm:

Yeah, and from a standpoint of developing and again I'm not gonna give tax advice on this, but when you're developing new processes, and especially when you're building new processes that are using AI and automation, these kind of cutting edge pieces you can either develop 100% internally.

Chris Tamm:

What we've found, and what our customers have found, is that can be very expensive because if you have one person on staff to do and it's just a recommendation that helps anybody if you have one person on staff that's responsible for building these types of tools, or maybe you have a couple, they usually go down a road and they usually can't understand, say, those 100 changes a day it's all the different tools so they usually get pot committed and start using a couple of tools. What we see as the benefit is we see all these cross industry projects and so we pivot people to different tools as new tools roll out, and so sometimes and it's not as a sales pitch, but sometimes it's easier to use a firm kind of like ours, because you don't have to rely on one person, only what they know. You get kind of the expertise of a group doing things across lots of industries, so you can move a little bit quicker, but it can still qualify under those R&D tax credits that's still being spent for that development reason specifically.

Mark Perlberg:

Yeah, and also when you have a consulting firm doing it, then 100% of those costs qualify as qualified expenditures for that credit, as opposed to another instances where we have people doing the internal research. They have to come up with a formula or a way to delicate what percentage of their costs and payroll is allocated to the process, development, the research. Here we have something that is clearly fully and substantially dedicated for the items that are our research and development. So can you tell us maybe some interesting projects that you, before we go into the tool, just maybe some things that, as a consultant, that you've helped to develop in other industry, other organizations that took AI and automation to the next level?

Chris Tamm:

Yeah, maybe a couple examples just from different areas. One that we see, I would say, fairly common is the idea of dealing with data entry. So when you have repetitive tasks so if we think of anything from healthcare where you have nurses or doctors or PAs entering tremendous amounts of information sometimes 100 pages of test results for one patient and those it's health related and so you've got to have a human typing those in or verifying. So it can be very expensive, time consuming, not super enjoyable for the person that's doing that. And so using machine vision and machine learning to have a system read those reports, even when they're very unstructured, and know where to put that information, where to put it in the database, and then in a sense give a person still as human in the middle that screen to see PDF or whatever came in, what was extracted, and they can just very briefly go through and review everything. And then using kind of the idea of an algorithm or other types of machine learning to even make recommendations from that data so that a doctor or a nurse can deal with recommendations, viewing, verifying, instead of actually having to type and think about everything themselves. And so for that we've seen like the idea of a human enhanced response or having a human in the middle, that it's not replacing a human but it's giving tremendous leverage and kind of flipping that 80, 20 of 80% busy work to then 80% patient facing. Another one, which is maybe similar to a tool that we could look at in a second here, is extracting information from pictures, or from pictures or video, so for an insurance process, for a tax process, for I mean, like when we talk about like real estate descriptions or properties, or analyzing qualities of something, or even getting property types and things information out of pictures where somebody doesn't have to do it.

Chris Tamm:

And then the idea of, like a sales manager, the three most common jobs that are getting removed right now from automation and AI. It's marketers, which doesn't remove the nuance of a good marketer to know how to grow a business, but the idea of the busy work that a lot of them need to do, like creating a brochure or creating a design. That can be done faster. But it's marketers, it's sales managers and it's developers. The sales manager one is interesting because a lot of sales managers they spend their time digging through data to try and find out what didn't happen, like what leads weren't called or who didn't follow up or what got missed.

Chris Tamm:

That's a great way for AI to kind of just know a system and then it can kind of just watch everything, whether through robotic process automation or other connections, and make suggestions like, hey, here's all the people that haven't been followed up with and here's all the messages we think you should send to these 10 or 10,000 people that you haven't followed up with or haven't contacted in six months, and so that we still recommend that you don't have a system, just go send all of those. You still have human in the middle. But the speed that a human can go through and just verify that messages are right. I mean one person can handle hundreds and hour, thousands a day and still have that human touch on it.

Mark Perlberg:

Cool. So now what I really want you to show and tell us about is the tool you have in the real estate space for creating listings, whether they're rental listings or listings for sale, so you have the ability to share your student as if you want to pull this up. But you showed me another time of this and this was like something that our audience would really love to learn about, because it takes a lot of time for them, especially some of our guys that have property business companies and managing tons and tons of rentals here. This would save them a ton of time, and I think that they would really appreciate looking at how they can use AI to evaluate these properties and create listings based on the photos.

Chris Tamm:

Yeah, absolutely so. Let me share a screen here. This is a pretty unique company because we see a lot of different implementations and we think of implementations of AI. Some of them are just I say just but a plug into the open AI and it sends chat GPT just plug in directly, so they're not really building anything on top of it. This is not a plug into another system. This is, it's called, my prop pal.

Chris Tamm:

I met the creators of this and I've spent a lot of time with them since. We've been to different companies in the real estate space, and this one's pretty specific because they have their own proprietary trained image database, so it's not just a plug-ins like Google Cloud. These are specifically trained on real estate and pulling real estate information out, and so you get much better information from it. The other piece is the text engine is trained around the real estate laws so it understands fair housing laws. So like if you went to chat to BT right now and said write a property description for this, that is not going to be fair housing compliant Unless you're auditing it and making sure that it is, whereas everything here is because it's built specifically for the residential real estate industry, and so a couple of things that you can do here. One of them is you can put information in, but I'm just going to put in a, just put in a house. I'm not going to type any features in, I'm just going to upload some images here and from those images, if we click, generate listing.

Chris Tamm:

So what this is doing is this is going through and analyzing every picture and you can go down to minute pixels. If you upload a video, it breaks down and looks at every single frame. So you could do a video, just walk through, of a house or of an apartment or have a tenant do it, and then upload it here and right away. What happened I mean, that was what a couple of seconds is it went through and pulled information out of each one of those pictures and it pulls out and it puts it into the 900 RISO standard fields, which is a part of the residential real estate industry in the US. So it can be ready to upload to an MLS, specifically if they're listing on an MLS, depending if that MLS has given access.

Chris Tamm:

But also it's writing the property description and it knows if there's stainless steel appliances, it knows if it's a tile countertop or a granite countertop, and then what it can also do is it knows a quartz countertops, maple cabinets, and then it can also make recommendations, like in this it saw a kitchen that is not fully up to date.

Chris Tamm:

Now what's interesting is I uploaded some random photos, but I did upload some pictures which was not an updated kitchen, and so it can make property recommendations on hey, you should do a fresh coat of paint to the exterior, you should update that kitchen with appliances, cabinets and countertops, maybe renovate the main bathroom with new fixtures, and so again, as you're showing it what to look at, I mean it's like hey, you might want to renovate this one. It doesn't have everything as nice, so it can look at quality, it can look at fair housing violations and it can write descriptions for you and then you can look at other versions of descriptions. So for us, this was one of those deep implementations where you can see right away that it's not going to chat GPT, but it's very much trained around a process. This company also has an API, so in a sense, an investor or somebody doing tax, I mean you could send a thousand photos past this and get back all the information you wanted and use it for your own process.

Mark Perlberg:

Wonderful, so yeah, so I'll probably take a clip of who's ascended some of our real estate agents. I think they really appreciate this. As we close out, some of the things I want to know are where are some of the things you do outside of fun?

Chris Tamm:

So I cut out there for about two seconds. Were you asking me personally what I do for fun?

Mark Perlberg:

Yeah, what are you doing outside of work?

Chris Tamm:

So it's either. So AI and automation is sounds super nerdy, but it is one of the things that I do outside of work. That's kind of what led it and bled it into work years ago. But either climbing, so bouldering, usually indoor out, hiking, paddle boarding up in Evergreen Lake just up the mountain up the hill, or hanging out with family. So pretty simple guy.

Mark Perlberg:

Awesome. And let's say, you give you to the audience one action item, one call to action. You have a chance to do it Now. Sometimes people can draw a brand, but I always like to give my guests a chance to give an action item, call to action for the audience and then tell us where they can reach you. So where are some of the action items you might want to give to our listeners?

Chris Tamm:

Yeah, this is probably again pretty nerdy and maybe going a little bit too far, but with how everything is changing right now.

Chris Tamm:

For me, ai and automation is not just part of what I do for work, but it's also understanding where the world's going, because when my kids ask me, hey, what should I go to college for in 10 years, this is a piece of my answer, because things are changing so quickly. So for me it's a real part of the world. That recommendation would be to spend five, 10, or 15 minutes a month and just Google what's changing in AI, and just even spend five minutes watching a YouTube video, anything to understand, kind of where this is going, because so much is changing, and the fact that a lot of people didn't know that ChatGBT existed until a month ago or three months ago, even though they could have used the same technology years ago inside of the Microsoft system, just knowing what's possible. And then, as far as how they can reach me, I think easiest would be LinkedIn, just christam, or email, which is christatchristamcom, or they could always take a look at our AI and automation firm, which is Cast Services. So cast C-A-S-T and then dot services.

Mark Perlberg:

Very cool, very cool, chris. Thank you so much for your time. I hope everybody enjoyed it. If you wanna learn more, you have Chris's info. It'll be in the show notes and if you're interested in being a client, email info at markoveredcpacom and subscribe for more wonderful, great content.

Chris Tamm:

Awesome. Thanks for having me, mark.