AI for Loan Officers: The Tools, Prompts, and Systems You Need to Know with Ben Tasker

In this episode of The Way with Dino Katsiametis, Dino sits down with AI and workforce transformation expert Ben Tasker to break down what AI actually means for business owners, mortgage professionals, and entrepreneurs.

Ben’s background spans AI research, data science, healthcare analytics, and higher education, including launching AI certification programs for over 225,000 learners and building predictive tools that improved real-world outcomes.

Together, Dino and Ben explore how AI can:

• simplify complex workflows
• automate administrative work
• unlock powerful data insights
• and give professionals their time back.

But the conversation also dives deeper into the philosophical side of AI — asking the question many people are wondering:

Is AI a threat… or the most powerful productivity tool ever created?

If you want to understand how AI can actually improve your work, your business, and your life, this episode delivers practical insights and real-world examples you can start using immediately.

Thanks for listening to "The Way With Dino Katsiametis"  
 For full show notes, links, and extra episode resources, visit dinokatsiametis.com.
Follow Dino for weekly clips and mortgage leadership insights on
instagram.com/dinokatsiametis | linkedin.com/in/dinokatsiametis
Learn more about Ethos Lending at ethoslending.com.  

Transcript

Din-Ben Raw: Hey Ben, welcome to the show.

Ben Tasker Raw: How are you doing today, Dina?

Din-Ben Raw: I'm doing great. Thank you. So I gotta tell you, Ben, um, everybody loves talking about ai. I love talking about it. I've talked to a bunch of guys in this mortgage industry that love talking about it, and, and I, I would even say that I'm pretty darn good at it now. And I have some friends.

We actually have this little AI fight club we call ourselves. But then when I read your resume, I am like thinking to myself, we are absolute kindergartners in this AI space. So I am super excited to talk to you because you, you actually have like this pedigree in, in ai, which I don't know how long, how far back that goes.

'cause I know AI's been around, it's just called like machine learning, I think, right? Like it's

Ben Tasker Raw: Status sign exactly.

Din-Ben Raw: But it's, it's just so much more now. So, so dude, if you don't mind, I just wanna get into it 'cause I have a ton of questions I wanna ask you and, and I think, [00:01:00] um, you know, you're not in the mortgage business, so you're, you're actually the first person on this, on this podcast that hasn't been in the mortgage business.

But I want to figure out how to take your experience, your knowledge, and then relate it back so that our listeners. Can say, ah, light bulb moment, maybe I should try doing this. Right? Like, so I ultimately, I want to give, I wanna give the listeners actionable steps on how to make their life a little bit easier.

Um, how to get better at doing things and, and how to get quicker at doing things. So with that said, I was, I was gonna like, introduce you, but then I was like, I can't, there's just too many things to memorize. So instead of me reading it, tell us who you are, man.

Ben Tasker Raw: So I'm Ben Tasker. I'm an AI and workforce transformation expert. I specialize in. Leading people through this turbulent, what I call ai. Between times I make individuals and organizations more AI enabled.

Din-Ben Raw: [00:02:00] Well, I gotta tell you, this thing that I just read is a little more impressive than that. I'm gonna read it. It's 'cause you, you obviously are humble and you don't, you don't actually.

Ben Tasker Raw: You're embarrassing me

Din-Ben Raw: all of this. So you're previously were the Dean of AI at Southern New Hampshire University. Where you launched applied AI certification supporting over 225,000 learners.

Ben is also taught at Northeastern University, Curry College, one of the top colleges for ai, AI and computer science. So I am just gonna read it. Uh, earlier in his career, Ben served as project manager at Northeastern's Experi, uh, experimental AI Institute. Leading AI solutions, designs and teams to help amplify human AI impact, which I really wanna dig into and find out what that means.

As a data scientist, um, at Maine Health, he developed, uh, predictive AI diagnostic tools, uh, that directly improved patient outcomes and saved lives. [00:03:00] And then there's more, but let's just kind of stop right there. 'cause, 'cause that is a lot more than. What myself and my buddies are doing. Like diving into ai, right?

Like creating things and doing stuff. You're, you're, you're doing it a much higher level. So let's start with what does human AI impact mean?

Ben Tasker Raw: Human AI impact is when AI can help. Amplify human beings. So, for example, some of the projects that I've worked on involve, uh, in the healthcare setting, there's what's called patient monitors. And those patient monitors collect live dynamic data and you can collect that data and create, create early alert systems.

Based upon the patients that are connected to them. So it's not really like a, a clinical triage tool. It's more of like, uh, the clinical team will get a page on their phone, like, oh, uh, Johnny's blood pressure is beginning to dip by a couple points. He could have an infection. You, you might want to [00:04:00] go check that out eventually instead of.

Reacting to it after it's already occurred, uh, is just one example of, of human AI impact. Uh, another example of it is in a, in a large school setting, uh, think about like massive online colleges taking algorithms to make sure students are successful. So if they haven't logged in in a long time, if they're not doing well in their courses, maybe they're in the wrong program, deploying resources so that.

Eventually they, they can become successful. The resources could be extra tutoring, it could be changing the course, the professor, whatever, whatever the course, uh, the, the remediation may be.

Din-Ben Raw: Got it. Okay. And. To the normal person. Um, AI has become a huge buzzword in the last, let's call it year, maybe two years for the people that were a little more ahead of their time. And then there was machine learning, which is really, it's still all the same thing, [00:05:00] but nobody ever calls it machine learning.

It's probably just not as sexy sounding as ai. But how long have you been doing it?

Ben Tasker Raw: Uh, it, so it depends when you want to consider what a at least 10 years. At least 10 years. So my own journey was an upskilling journey. I graduated from my undergraduate degree. I thought I wanted to become a healthcare administrator. I went into a ha. Hospital and one of my first tasks was to observe patient volumes at different clinical care offices, and that this is another example of a human impact project, and that what ultimately that means is that they might shift around some of the practices based upon the volume.

I thought, oh, no problem. I can just go get the data. We can make some simple charts and graphs, and what I learned was sometimes the data didn't exist. Sometimes it was still on paper, and this was not that long ago. It was on paper, not in a database. So I had to physically go and talk to the clinical teams and really develop those human skills so that [00:06:00] we could gather the data.

Some of the clinical teams actually didn't really wanna share the data because they knew of some of these projects going on. And that's when I, it really clicked to me that that data's important. Data-driven decision making is important. And around that same time, data science master's programs were starting to emerge.

I already had the business background getting an MBA. I'm more innovative. I like playing with things. I really love the data. And I was like, why not do data science? It combines all the stuff I like, and that's where I got to learn about the AI algorithms that I apply today.

Din-Ben Raw: Yeah, I love it. So, I, I gotta say, I'm a, I'm a, I'm a data geek. Um, but I'll tell you where my biggest frustration is, is the lack of data. So to, to specifically speak to the mortgage industry. Uh, you know, there, there are maybe some of the bigger companies that have more data and they're able to do stuff, but it drives me crazy that. [00:07:00] A typical loan officer slash mortgage advisor, um, they, they really don't maintain data the way they should, and therefore, I would say it's impossible to even analyze it, uh, properly when, when you don't have enough of it, right? So, for example, um, how many referrals did I get from that real estate agent? How many referrals did I get from that financial planner and that attorney and all that? And, and man, I, you know, I got so many more from that real estate agent that that's absolutely my highest and best use. I'm making the most money, right? But, but the reality is, is it really meaning how many people did they refer you?

How many people did you call, uh, or how many people, how many of those referrals did you actually have a call with? How much, how many resources were allocated to it? How much time did you spend on it, how much money did you [00:08:00] spend on it? Pre-qualifying people, 'cause there is a cost associated to that. And, and then ultimately, what was the, either the, the, you know, uh, dollar made per hour or the dollar lost because of all that you were doing.

So then we could take another sector, like the financial planner maybe, who

Ben Tasker Raw: You are kidding. You really do love the data. No, I'm just kidding. Keep

Din-Ben Raw: Well, it drives me crazy 'cause we just assume a lot of things. Right? But, but when you have all of that and then you have the other data and maybe you get less referrals, but everyone closes, you gotta kind of ask yourself, which one did you make the most money on?

And, and people don't realize that, like, they don't dig enough into the data. So for me,

Ben Tasker Raw: Well, it's hard to set up too, right? Like everything like that's. It's not super complex, but you still gotta put the time in to do it right.

Din-Ben Raw: That's the problem. You gotta put the time into it. And, and I think for the most part, even a lot of the bigger companies, [00:09:00] the loan officers are, are very solopreneurs. Meaning they, they function off their cell phone, you know, still, they, um, they don't properly use a CRM. They don't properly take the time to, um, you know, to, to notate everything that happened the right way.

And, and then even when they do, there is no proper, you know, data driven analysis afterwards to really determine the value of it. So, I, I, I am on a mission at Ethos and, and I've, I'm gonna just say it out loud. I've already failed several times. Um, which drives me insane. I hate that. But as they say, you know, every time you fail, you get one step closer to succeeding.

So, so I've, I've failed multiple times, and the reason why is because we use so many different systems that it's hard to like round them all up and then [00:10:00] pull from each one. I, I am hoping now with all the AI stuff out there that it will get a little bit easier. And, and maybe even doable for once. Um, but, but that's my mission right now is to, to pull all this data, have ai, analyze it all, and be able to spit back real answers and hopefully even real time to, to really, I, you know, really make a, a, a proper diagnosis of what the next good step should be.

Ben Tasker Raw: So a, first of all, ai, just like you loves data, everything that you just explained for a minimal subscription to any AI platform or maybe the top three AI platforms can help a lot of mortgage brokers like yourself become more data driven, collect data, create a mini customer service management system.

Uh, it's all right there. You just have to properly label things and [00:11:00] use the technology in front of you. So, for example, with. An iPhone. I'm an iPhone user, or, and I know Google also. Google's Pixel also has this option, but when you make a phone call in the top left, there's three dots. You can click those three dots, click record and transcribe.

Once you have permission to do that, the full call, since you're already in the call, your your name will be printed and assuming you announce the other person's name, that will be printed so that the transcript's very easy to to use. It will record the whole call. You can then copy that transcript. It goes right to your notes.

You'll, so now you don't have to type anything. You can label that file in your notes. Uh, onboarding call with Ben Tasker, first time home buyer, X area, or whatever the call was about. You have that. Then you click the three dots again, you can copy the transcript, or you could just put the file in, but you can copy the transcript open your favorite AI tool.

You start a project. [00:12:00] Assuming you already have a prompt set up and, and you have the project going, you can put in the transcript and it will fill out your customer service template. Some of the things on the customer service template will be the person's name, contact information, the date you contacted them, what you chatted about, follow up items, it can even draft emails to follow up with, help you with, uh.

Setting the reminders. So you'll still have to program them or put them in your phone, but you can just copy paste them. Um, but that's all at your fingertips now. Uh, I tried it before we had this call, and it took 15 minutes to do all that. I am not a mortgage worker. I don't know how long that would take without ai, but from a six minute phone call to all the steps I just described, 15 minutes total.

Din-Ben Raw: Love it. I'm glad you brought that up. So it's a pet peeve of buying. I mean, I, I've stopped working on my cell phone a long time ago, but. Pretty much every loan officer I know works on their cell phone and, and they don't [00:13:00] capture any of that stuff, right? And, and I, I get it sometimes recording a phone call, it's kind of weird, whatever, but all you really need is the transcript.

You don't need that, the actual, you know, anything. And, and then I love what you said, like it took 15 minutes, so I, I believe you should schedule. 10 to 15 minutes after each and every phone call you have just to properly label it, set reminders, put in everything you need, and then it's just done after that.

And in these transcripts, I mean, I, I actually, so I use a, uh, a platform called Rome. It's, uh, it's like a virtual headquarters for, for ethos lending. And, um, they just, they just rolled out a new, a new deal that, which I love. Right. So. I tried doing, I tried just staying off the phone. Now I tried doing everything on, on virtual.

If I get the video, great. If not, it's still being recorded through there. [00:14:00] And, and they just actually let us set up these templates. So if I'm doing an initial consultation, I can set up one template, uh, for a, for a buyer, right? If I'm doing a interview to hire a loan officer, I can set up another one. And like the, I, I focus on, on recruiting.

Uh, and I started playing with it. And what I did, I did my first one yesterday. Um, I set up the template so it, it tells me like the scenario or the synopsis of, um, who the person is, what they're looking for. And then I have another little piece, um, down below that says like, what their pain points are at, at their current, you know, in their current environment.

Um, I have another one just below that that says what they're looking for. Um, and what their like dreams and goals are. And then I have the bottom one that says, analyze their disc profile and gimme a confidence score with it. And, and then I had my, my first one yesterday after I, you know, [00:15:00] set up that template and it was so cool.

Van, it was like so cool to like, have it broken out for me exactly how I wanted. And I, I, what I did is I focused solely on my conversation. I didn't have to take any notes. I didn't have to worry about anything. And, and this is the point, I think what I'm so excited about, if loan officers can start doing this, anybody, every, everything, and start trusting the system, you can spend more time actually connecting with somebody.

In trying to have a more meaningful conversation. And then what I love about it, if you take a minute to go through the transcript afterwards and break it up like this and, and actually ask it questions and get some feedback, and boom, boom, boom, will, you will learn for your next conversation of what you didn't ask, and you'll start having a better conversation with people because you're no longer taking notes, but you know it's there.

And, and if you don't [00:16:00] ask it, it won't know it, right? So start asking better questions. Start asking more, and, and stop worrying about trying to write it all down. I'm talking about a loan officer perspective, right? Like, oh, you know, where do you work? How much do you make? You know, it's like, ask those questions, it's fine, but ask it in a more conversational way now as opposed to a, a, a loan application being taken, right?

Like, so much better. So I'm glad you brought that up.

Ben Tasker Raw: Yeah, absolutely. And I'm sure that mortgage brokers also have their own document process, so even on these onboarding calls, you can be more interpersonal, you. Know the person. It's not so much about getting every little detail. You could even send them the document, they fill it in, and then you easily digest it.

Right? And you're gonna probably have to do that anyways. And that fills in your other templates. So by thinking about this at scale, right, th this process might take you longer in the beginning, but throughout it you're probably saving, I don't know how long it takes for the complete process, but probably a couple hours and you're getting better customer service, [00:17:00] which probably leads to more closures for you.

Din-Ben Raw: Um, I gotta figure out how to phrase this question to you. I am going to make a comment 'cause I think that'll be the question and then you can answer. When I started going deep into, um, trying to learn AI to be more than a glorified Google and really spending time with it, I, I, my wife noticed, um, how much effort I was putting into this. And instead of being like.

You're such an awesome husband and I love your tenacity on, you know, wanting to learn and blah, blah, blah, blah. Instead she was like, why don't you just ask your girlfriend? Or she would say, little comments, and she was getting sick of hearing me talk to the computer or my phone the whole time. And, and then I got kind of excited and I started trying to teach my 13-year-old a new [00:18:00] way of studying for his science exam that he was struggling with.

And I'm like, Hey, let me teach you a brand new way of doing it. And I, and I basically taught him how to put it in Notebook, lm and upload all of your questions and answers. Upload, you know, a couple other things and, and then, you know, watch this video. 'cause he's trying to look at a book. And he's not getting it.

So it upload it. Boom. And for all you guys, let me back up for all you guys listening right now, if you don't know, but what Notebook LM is, um, it's a Google product and if you have Google, you have Notebook, lm, and it's one of the coolest things. You, I literally took pictures from this textbook. Um, I uploaded them there and then I even put in just a little thing of, you know, this is my eighth grade, you know, son.

Um, he's trying to study for this test. He is struggling with it, help him to come up with a, a better way, right? So it is something like that. And then if there's YouTube videos or if there's websites, you can put [00:19:00] those in as well. And then this AI agent just creates, and this is what it did, it created a, a, a 10 minute video explaining everything that was on his new test. In an environment these days where kids watch YouTube, like, you know, like it's going outta style. Uh, they, they actually learn from watching videos. Well, you know, science teacher doesn't teach with videos. They just have this textbook and a few, you know, it's just not the same. And then it was super cool because it creates a podcast of two people talking back and forth explaining how the, you know, the duck,

Ben Tasker Raw: you were gonna bring up the podcast.

Din-Ben Raw: What's that?

Ben Tasker Raw: knew you were gonna bring up the podcast feature.

Din-Ben Raw: Yeah. I mean, and it's explaining how the duck can be in, in moving water with swells and all that, but yet not moving with the swells. Right. And, and the current underneath and all this stuff. And then I was like. Watch son. [00:20:00] Uh, and I hit the button on the podcast while they were talking. And then you have these two AI people saying, oh, hold on, we got a, we got a question from, you know, from the listener.

Uh, and, and I'm like, I still don't understand why the duck doesn't move along with the swells and great question. Let me explain it again. And it just goes right into it. I was like, this is amazing. Right? So

Ben Tasker Raw: That's three, maybe four types of learning too, right? I know you were building this for your son, but your son could do it eventually, right? So that's one type. Second type's the video, third type's listening, and then you kind of have to confirm if it's true or not, which is a higher form of thinking. You might not be able to do that when you're first learning, but Right.

Like you have to validate it a little bit. Right.

Din-Ben Raw: yeah, yeah.

Ben Tasker Raw: that's probably way better learning.

Din-Ben Raw: and then it created flashcards. And, and, you know, and he was actually just like loving it. He was so enthralled with the whole thing that he really took to it. And, and he wanted to make sure he got the flashcard right, right. To [00:21:00] him. It, it, it

Ben Tasker Raw: Make it a game.

Din-Ben Raw: Make it a game. So that I went into pretty a, a pretty big story there.

But my question was, my wife saw all this and, and her being her little conservative, uh, Midwestern girl, nature said. This is the devil, and you're not doing the kids any justice. You're, you're teaching 'em to just be solely dependent on, you know, this machine, which, you know, if Christian terms right, the machine is, is the devil.

Um. I'm like, it's not babe. I'm like, it's not the devil. In fact, if anything, I think it's God saying, let me help make your life a lot easier so that you can actually have more time to spend with people. Again, more time to spend with your wife, more time to spend with your kids, more time to like enjoy breaking bread with friends and not being so like tied to every little thing that's happening over here.

Right. So I guess my question to you is. Is [00:22:00] AI the devil or God?

Ben Tasker Raw: Ooh,

Din-Ben Raw: way of asking it. You know what I mean though?

Ben Tasker Raw: I do know what you, I do know what you mean. This is a fun question. So my fiance actually kind of had a similar conversation with me because I thought it was so cool to create. You can do movies, you can, it was so fun for me, and I'm not super like artistic, but to be able to go make a movie that I would never be able to produce myself.

It's really cool to me. Um, so yeah, similar conversation. And she's also a psychiatrist, so she's seen some impacts of, of AI from a patient perspective that I maybe don't consider when I'm, I'm doing some of the cool stuff. Uh. I think with trade offs, like for learning, you can obviously get some time back.

For a mortgage broker business, obviously, no brainer. Why would you not make your life more simple? But then there's other cases where I've heard people having, or trying to have physical relationships with ai, like it's, it's a partner. Uh, they overuse it for empathy. Like [00:23:00] it gets so personalized. It's actually not treating them for anything.

It's just telling 'em one, enforcing bad behaviors essentially. Uh. So you have to know when not to use the technology, but think about other digital tools, internet, social media, those had some negative impact, especially on children. If you can pull 'em out of that and they're learning about ducks, right?

Like maybe better, maybe better. So I think it's mostly good to answer your question mostly.

Din-Ben Raw: I, I agree. Uh, I think obviously. There's, uh, you can, anybody can find bad in anything, right? It, it's just how you're gonna use it and all that. But I guess where my thought process is, my, my goal is to do 12 hours worth of worth of work in six to eight hours. Like that's my personal goal. And no longer have to put in the long hours at the office.

No longer have to like stress out over everything later, even. [00:24:00] Even, you know, when you're in sales, it and, and in and in our, my world. It's not banking hours. Banking hours is when you work at a bank and everything shuts down and nobody cares after it shuts down. Right? In my world, you don't get a check until the deal closes.

So, so everybody on my side cares. Doesn't matter what time of day it is, they care. And I just think that with proper, um, AI use and setup, you can have full coverage all the time, even when you're sleeping. I don't know exactly how yet, but I do believe that the majority of the little things that happen after hours can be taken care of through ai.

So that's, that's my, my big goal, and that's where I hope that it goes with everybody. Now, of course, if somebody wants to be a workaholic. They're gonna, they're gonna take, you know, 12 hours and shove really like a 20 hour day into 12 [00:25:00] hours. So be it. If that, if that's what they wanna do. Right. Like, I guess my, in my old age, I'm, I'm trying to get to, to the, the less hours instead of more hours piece.

Ben Tasker Raw: Yeah, so, so AI can definitely unlock those opportunities. I think it's oscillate. You have to understand the risk too. So, so bringing it back to the, the customer management system. So the next step in that is, is creating a project. So most AI tools have. A project category where you can have different folders for you.

It could be the different templates that you're using. They'll have what's called system prompts. It's a series of steps that the AI will follow once you load your transcript into those, those files now. The AI tools themselves can't compose. Like they can compose the email, but they're not attached to your email, so they won't send the email and they, they can't schedule your calendar invites, but you can connect them to applications like NAN, uh, Zapier that do have access to [00:26:00] that information and can do that for you off hours if you want to have it done off hours, but.

You also still have to check it. So maybe instead of having it fully composed, the email, saving it to your drafts and then checking it to make sure that it's accurate and, and it actually matches the transcript. 'cause sometimes I've had it where it totally screws it all up and it's not anything I would look on and say, or it schedules the, the calendar invite for the wrong day and I miss something by a whole day.

'cause. It's ai, right? It's not perfect yet. Uh, so there's still some intervention there, but yeah, you can auto, and those are very simple things for someone, right? Like probably scheduling could be a couple hours of your day if you don't have an admin. Or if you do have an admin, more help the merrier, right?

Social media, for example too, it can help you create posts, it can design images for the posts. It can suggest where to post them. It can give you the words to put in it. You still want to make sure that it's right and it's structured correctly because AI, [00:27:00] again, is a recommendation system, so it's recommending something that it thinks you wanna see.

It doesn't necessarily mean that you wanna see it. Uh, I actually just tried asking AI to give me a random number one through 10, and I've, I've tried it a hundred times since. Uh, most times either giving me a seven and then a very small amount of times giving me a three. So it's deter. Even though we say it's non-deterministic, it does have some determination in there.

Uh, it's a magic trick. Try it. I I bet you you'll get a seven or a three, uh, if you ask it to give you a random number one through

Din-Ben Raw: That is so interesting. I'm gonna do it right now.

Ben Tasker Raw: see, let's see, let's, let's test it.

Din-Ben Raw: Yeah. Can you gimme a random number between one and 10? hold on. Well, I don't know if my internet's going bad or what. Let's try this again. Can you gimme a random number between one and [00:28:00] 10?

Ben Tasker Raw: I'm.

Din-Ben Raw: I'm going with a bold and optimistic seven. Lucky

Ben Tasker Raw: I

Din-Ben Raw: keep things

Ben Tasker Raw: didn't need to try.

Din-Ben Raw: not too high to scare anyone off. Okay, let's try this again. I want you to gimme another random number between one and 10. All right, let's shake it up. This time it's a three. A little underdog number, but it's got potential. All right. There you go.

Ben Tasker Raw: So non-deterministic. I mean, that's a pretty impressive magic trick. But why does AI do that? Is because it takes all their written information, Reddit, social media posts, whatever the case may be, that it web scrape. And those are the most randomly selected numbers in that range. So it doesn't actually calculate anything.

It just gives you the number that it thinks, thinks you're gonna pick, right? Like it's not even random.

Din-Ben Raw: I try a little test real quick. Um. Hey, just for fun, pick a random number between one [00:29:00] and 10, period. I'll tell you why. When I get home, I sent it to my 13-year-old. I wanna see what he, what he chooses.

Ben Tasker Raw: Right. Yeah. People are like, how'd you guess? How'd you know it was gonna do that?

Din-Ben Raw: Uh, that's so funny. All right, so let's, uh, I want to go to another, another question. Um. Platforms. There's chat, there's Gemini, uh, perplexity. Um, I don't, I don't, there's, there's, I'm sure more. Uh, my son shows five, by the way. Um, so let's talk about platforms for a minute. What are your thoughts, Claude? Um, what are your thoughts on which one and what are your, I guess, top reasons why? And then also frustrations with them.

Ben Tasker Raw: So chat, GPT opened first, full disclosure, pretty much the second it launched. I was in there. I saw the potential of it. Uh, [00:30:00] once you're in, it's kind of hard to get out. It's a preference in that, that sense. It knows me, it knows, it knows me really well, and I have all my projects in there. Um, Gemini's also really good.

It's connected to more of the, the Google Suite. They also have what's called Google Gems. It can make basic applications for you with a prompt. Super fun. Um, you can, you can mess around with that. Per hours. Claude Claude is considered at, at least right now, one of the more intelligent artificial intelligence.

Usually it's 2% higher because of how they do their modeling technique. Um, it's great for coding as a data scientist. Claude Code is. Has saved me literally hours, and in some cases, does the work better than I would do it. Um, it it's very, very impressive. So in, in that case, I kind of think it depends on what you want with the features.

I think some of the trade-offs are knowing what you want to do and which tool to use. Uh, think about it like a skill, right? Like, oh, Claude's really good at coding. I can kind of relate to [00:31:00] that. I already have it set up. Maybe I'll use, use Claude for that one. For the customer service management system. Uh, I, I went into GPT first, then I tried it in Gemini just to see if I would get a similar result.

Um, but I have my GPT already set up, so, so that was kind of why I went in that direction. The frustrations with it is GPT now knows me so well that I have to change my profile information, so it's not. As personalized, um, or, or gives me the same responses every time. And you can just kind of tell that, and I'm sure the other tools will do that with overuse as well.

Um, but the way you can tell that is that everything kind of just reads and sounds the same. So it's, it's time to switch it up with, with cloud and cloud code, you kind of have to check it against something else. So I actually check it against GT's Codex to see, see if the output is correct. Quote unquote correct.

So using them against each other sometimes helps.

Din-Ben Raw: sure. Um, so [00:32:00] I'm with you. I, I, I use chat, uh, because I've got so many hours into it that I can't even fathom the idea of switching. But I will say, you know, I switched my company's entire platform from Microsoft over to Google because of everything happening in AI and because of all the products Google has and, and, and yet, I still haven't switched over.

I used Gemini to. To like polish my emails all the time. Uh, and it's great. Saves me a ton of time. But the other day I was, you know, I did the whole transcript thing after a call and then I tried to, to get chat to, you know, and I, I got frustrated 'cause I'm like, great, set up a calendar reminder for me, you know, to follow up on this day after it told me to, to do that.

And it's like, oh, I can't, I can't quite do that for you just yet. I'm like, why? Because you, you asked me to connect my, my

Ben Tasker Raw: with the application,

Din-Ben Raw: and my Google drive. [00:33:00] Why can't you do it? And, and it drove me nuts, right? So then I just went over to, to Gemini and I'm like, Hey, set up a calendar reminder for me.

It's like, no problem. And it just did it. So I started thinking about it. I'm like, am I being a fool? Should I just get started before I have any more hours into it? And can't, can't switch, right? Like, and, and, and then I kind of wonder, so here's my question to you. And I think I know the answer because, 'cause you're doing the exact same thing.

You're so, you're so entrenched, you won't move. Uh, which by the way, your fiance should be super happy about that. Like the chances are you'll, you'll never leave her. Uh, which is great. But I'm wondering if there's a way for me to take what I've done in chat and, and capture it somehow and then upload it to give Gemini a head start.

On who I am and everything I stand for and all my values and all my core [00:34:00] beliefs and um, and everything that I've done, all the countless hours with as smart as AI is now, I mean, it's way smarter now than it was six months ago. Can't I fast track it and not really lose all of that? And is there a way.

Ben Tasker Raw: I'll answer the questions, you know, but do you wanna do that? Because then you're also carrying over the biases, right? Or potential biases maybe carry over some of the processes, right? Your customer service management or something like that. But I don't know, would we, would I want to co convert everything over some of the stuff I do in there is just for fun.

I think it's important, but it's, I don't know, it's kinda like stuff in my house is probably not that important.

Din-Ben Raw: yeah,

Ben Tasker Raw: so

Din-Ben Raw: that's a good point. I, I'm thinking that if it can add things to my calendar, for me alone is

Ben Tasker Raw: your team,

Din-Ben Raw: a really. Positive thing and what,

Ben Tasker Raw: And your team, right? Like you're thinking of Right. Scalability. I mean, if that's the future for you, that's worth it, then you [00:35:00] could convert that all over and, and that would be, you'd take your system prompt. I'm assuming you have all your transcripts. You can put that over there. You would set it up the same exact way.

Gemini doesn't really have projects though, so they'd be individual chats. Um,

Din-Ben Raw: gems, aren't they?

Ben Tasker Raw: you can, yes, you can use the gems, but it's not this, it's similar but different.

Din-Ben Raw: Okay.

Ben Tasker Raw: You, it can build an application for you that does. Something similar, but do you need an application? I'm not sure.

Din-Ben Raw: I, I guess I feel like I, I was being a fool because I was like, oh, Google's gonna, Google's missing the boat, Google's this, and then all of a sudden Google is. Not only, you know, stepping it up big with Gemini, but I didn't like it at first all that much. Right. Compared to chat. But, but it does seem to be doing pretty good now.

But then when I, when I dug into like Google Labs and all of the, and, and Notebook lm, and I'm just like, oh my gosh.

Ben Tasker Raw: vo, I mean, Vos [00:36:00] great for videos. I mean, I think it's superior.

Din-Ben Raw: It's awesome, right? Like all under one platform. Uh, and I can't help but wonder, would all of that make my life so much easier in the near future? Or should I just chill out and wait because another three or six months, who knows what's gonna be there?

Ben Tasker Raw: Fear I'm missing out. Right? Uh, so you can use your other AI tools, you can con connect them to tools like NAN Zapier. You can get all the same functionality. Yes, it's a little bit more expensive, but if you have everything set up, the time it takes to unset it up and put it somewhere else and try to get similar results. Um, so they're all work, work around with that. Um, I have access to all three tools, so I use them interchangeably. So, um,

Din-Ben Raw: What do you have access to?

Ben Tasker Raw: I have all three of the main tools. So I have Claude, I have Gemini, and I have GPT. And I, like I was explaining before, I use them a little bit [00:37:00] differently. Um, depends on the task, but I, I don't think folks should be afraid of moving out or moving in.

I think folks should be afraid of actually not using it.

Din-Ben Raw: Yeah. Okay, so, so let's give some advice right now and let's give it to loan officers. They know about it. They think they know it because they use it to ask some questions. Um, or write a couple things, but, but they don't, you know, uh, they don't have a clue to the depth of the capabilities, but they really need to before they fall so far behind.

What is the first step? How do you start crawling before you start walking and then running? What, what, how would we give somebody advice to, to jump on this bandwagon?

Ben Tasker Raw: I started, so I. I knew the capability, so I [00:38:00] can't say I didn't know, but I didn't know the full capability in which I'm seeing now and it gets, it gets better over time. I was using it for fun and hobbies. I started using AI to plan my workouts. I would take photos of my workout equipment and then have it plan me a workout that, that I should, should suggest I should do, or my meal plan.

That got me.

Din-Ben Raw: even think of that.

Ben Tasker Raw: It got me to trust it. It got me to trust. I'm like, can I trust this? I don't know. Um, I was really apprehensive, but I, if it's a workout, so you can't, it's, I'm gonna do this anyways. And I don't, it reduces cognitive load, so why not have the AI try to plan it? Um, meal plans. I can then link the meal to the workout.

Maybe get better results. Um, that was fun. After that, how can I use this for work? And then I've started thinking about the skills. So coding is a skill, not just like data science, but actual coding, which is a little bit different because you have to be able to do different codes and different code stacks for that.

Um. How to interact with people. Now that, [00:39:00] that's a weird one, but you can, like you were doing chat with any of the tools and over time it can give you feedback, it can help you prepare for an interview. It can, you can just talk to it if you want to while you're driving. It sounds weird, but sometimes that's nice to do if you get bored.

Um, so it's thinking outside of the box a little bit and then as you get more comfortable maybe putting in your business and, and whatever else in your life that you want to include.

Din-Ben Raw: Yeah. Um,

Ben Tasker Raw: Learning doesn't have to be academic, I guess is what I'm trying to say. Like you don't have to go take a PhD level course to learn this stuff. It's, you can get as much or little as you want out of it.

Din-Ben Raw: Yeah, and, and I think I would say that if somebody doesn't know. They should probably start making an effort to go to, whether it's chat or Gemini or whatever, [00:40:00] um, instead of Google just to ask a, a, a simple question, right? Um, maybe that's a good way to start. Now you're just glorify Google user 'cause you're getting a little bit more out of it.

But then I would say switch to something fun. I, I love like your idea of like the workout. Just take a picture of your gym or, you know, if, if, if you have a home gym or something. The other thing, and I've never actually done it, but somebody said, oh yeah, I just took a picture of my refrigerator, um, opened the doors, took a picture, and asked it to prepare a meal for me to gimme a recipe.

And you know, with everything I have in there, I'm like. You can't even see everything in my fridge 'cause it's so stacked with stuff. But, but kind of a cool piece. Another one was, after your meal is done, take a picture of it and ask it to give you the, the macros. If you're into that. Um, as well as like one of the guys was like, uh, inflammation guy.

Like, uh, I can't eat anything that has, you know, that's gonna gimme inflammation and gimme a inflammation score. It said, and, and you know, he was like, all right, [00:41:00] can't eat that, or I need to take this out. And, and then he started getting into better eating habits because of that. Um, so those are all fun things, but then I would say get into something more fun.

And I'm talking work-wise now for loan officers. Uh, most loan officers are on social media. So try using it for some social media. Um, have it write up, you know,

Ben Tasker Raw: At this point, it knows you now a little bit too, right? So it knows you a little bit. Sorry. Keep going.

Din-Ben Raw: No, no, but you're right, right? 'cause it'll start to know you, you'll get to know your personality. Um, you know, uh, maybe look at and, and just ask it.

Even you can, you can find your local news yourself. Um, or ask it to find the local news and then relate it to something in mortgage so that your clients, um, would show interest in it, right? Some, something like that. And just see what it does. And then boom, write yourself a little article, um, you know, ask it to create a picture, a fun picture.

And, and that's I think, just a really good way to start [00:42:00] where you can start thinking highly of yourself because you got it to create a picture and you got it to write a little. And, and then all of a sudden you're like, well, let's try this. And then next thing you know, you're, you're in the 10 foot deep zone and you're like, oh my gosh, I can do so much more stuff.

And. I'm gonna turn around and say, everybody make a huge effort to start recording every single interaction you have with clients. There are, there are lots of tools. You know, Ben was saying, just on your own iPhone, you can hit the little button. Um, you can, you can have tools, you can have Fathom ai, you can have fireflies, you can have Otter, which is the one I use, and if I have a live conversation, um, I can still hit the button and it just records everything and then uploads the transcript.

And then, you know, analyze, you ask it, analyze it for me, ask it to predict certain things, ask it to, uh, analyze you. Um, and what could you have done better? You know, and, and it like, I did it and it said [00:43:00] you really connected great with the client, um, when they showed fear of whatever. Um, and then a little bit later it said, but you could've.

Talked a little, expanded a little bit more upon this instead of blowing right by it. And, and I was like, huh. You know, I, I guess I could have, right? So my, my recommendation to everybody is, and, and this also satisfies my wife's, you know, issue with it, of not becoming, she wanted the kids to become a critical thinker, like on their own.

And, and I said. I'm a much better thinker now than I was six months ago, but I don't just read the first paragraph of what it says and, and be like, oh my gosh, this is awesome. And then, and, and put it away. I read the whole thing and I'm like, Hey, this isn't exactly true or the way I would say it. Make sure you update that or, or add this and make it a little bit more of this tone so it sounds [00:44:00] more like me.

Even though it knows me, it's still. It still does these things every once in a while, right? So make do yourself a favor. Read every, every ounce of whatever it wrote, and, and then don't be afraid to question it. Don't be afraid to spar with it. In fact, even ask it, spar with me. Give me, gimme five things that you don't like about, you know, what, what I just put together or something.

Right? Let's make it better.

Ben Tasker Raw: You can do so much in the tool too. Even with your images, you can click on the images and highlight areas that you, you don't have to reproduce the whole image. You can just produce parts of an image and it's different types of prompting techniques. You can ask it to remember specific things, not to remember specific things and, and it will try to carry that forward in later conversations or images It.

The next level up is then creating a, a system of prompts so that you can start doing tasks. And, and sometimes even though not [00:45:00] everybody's gonna need to do this, but trying to create a, a video, I mean, it doesn't sound as hard as I'm explaining it, but the prompts you need to do for that, if you can create a video that you would actually wanna watch.

You can create any tasks in, in any of the AI tools for, for your business. Uh, it's actually easier to do that, but sometimes for me, I had to do something harder to be able to understand how to break down the system level prompts for each individual task. So by playing,

Din-Ben Raw: talk about prompts. Let's talk about prompts. Um, I will say that I, I could get better at prompts. I'm, I'm a lot more conversational and, and sometimes I wonder if that's a mistake, if maybe I should be a little more concise and, and get better at prompts. What, what's your suggestion there?

Ben Tasker Raw: So going back to the customer service. Tool agent, potentially, whatever you want to call it. Um, this is the prompt I use and, and then I'll explain it [00:46:00] again. I don't have any mortgage broker experience. I didn't have a template, but you could, you could have that. I pretended like I did it was able to produce it.

But anyways, here it's, this chat's going to be for a mortgage broker customer management system. I'm going to use my phone to record and transcribe client calls with permission. Your task is to create a CMS template from the call transcript. You'll help me draft follow up emails and conversations. So there I gave a pretty good context of the character it needed to play.

AI loves taking on a character. It can be anything in the world, and it really helps narrow down its knowledge so it won't hallucinate. I told it exactly what I was gonna do. I didn't drift off. Uh, it one sentence right there and then I asked it. To design its own template, which it can do 'cause I'm sure other people are doing this so it knows.

Okay, mortgage broker, simple task list, that's easy. Um, or I could attach my own too. And then I'm drafting conversations and emails so it can produce that. As soon as I put in this [00:47:00] transcript, I don't have to re-ask.

Din-Ben Raw: Awesome. Um, another question for you, a friend of mine, um, and somebody I interviewed on this podcast, he's now a friend. Um, his name is Eric Post. He owns a company called HU z.ai, HUZI. Which Eric, if you're listening, I love the name. I still love that name. Uzi. Anyway, um, he has something that re, that I struggle with, but I, I think I understand the importance of it, so I'm trying to kind of force myself to, to use it.

So, um. He has like the regular AI chat, GPT. You ask it a question, it tells you all normal stuff and you can choose if you want Gemini or if you want, you know, GBT or if you want Claude, you can just choose and it'll answer all in one platform. So that part's pretty cool, [00:48:00] then he has this other, you kind of toggle over and it's called Canvas and I don't, I don't even know if it's open to everybody or not,

Ben Tasker Raw: I'm on the website. This looks pretty cool. Keep going. Sorry.

Din-Ben Raw: So, so Canvas, you basically start with this blank canvas and, and you throw up a like little, we'll call it, you know, Chachi PT window and, and you can talk to it. It'll act just like Chachi pt, but, or Gemini, whatever, right? But then you can click another button down below and it'll throw up a, a YouTube link.

You throw in a video from any YouTube video, you just put the link in there and then basically it'll read everything that's in that video and dump it into the chat GBT box. And then you can take a book, um, and, and put the book over here and then draw a line into the chat GBT box. And then you can do like all these different things.

So you can even. Take like, uh, let's call it Tony [00:49:00] Robbins, um, as a, a, a personality or a persona I think they call it. And, and now you can ask it to create a social media post, um, or a, a blog or whatever based on these three things. And, and I'm like, I guess you can just write that into chat. But at the same time, if you have a very specific video.

Um, and you throw that link in there and then you throw in the, the personality of Tony Robbins. And I want it to be very, very motivating. And then you throw in this book because it's the values of whatever. And his, his concept there is that too. All too often if you try doing all of that in one chat box, it does get a little overwhelmed and it does start to drift a little bit.

Um, and it does start to like, kind of come up with its own stuff. Versus if you limit what goes in, then you get better, better results.

Ben Tasker Raw: This is, this is also workflow [00:50:00] optimization too. Like there's other tools, but by having it all in one place, you don't need a bunch of different stuff. So that's, that's cool too. Right?

Din-Ben Raw: yeah. And then, and then the other day he threw something out that I thought was really interesting too. All on this one canvas. You can, um, have all of that stuff we just talked about and then you can have like your chat GPT box where everything flows to that and you're creating whatever it's, you're trying to create.

Then you can have the same three items that we talked about, the YouTube video, the book, and the personality, and then have another box only. Now it's your Gemini box. Then you can have another one down below, which is your Claude box. So you can do all of, all three at one time based on this, these, the, the, the three things that you've connected to it, and, and then you can have another box that takes the best of the three and now works here.

He said, when you try doing all of that, it, there's, it, it is just too much. [00:51:00] Um, and it gets way too confused, right? And starts to, to drift and, and hallucinate and do all sorts of stuff. But if you keep breaking it down into these little aspects and then you put it all together at the end, it's very concise as to what goes to the, you know, the main, the main board.

Ben Tasker Raw: I'm visual too. So by having all of this visual, you can kind of, you, you actually definitely understand where it's getting that output from, right? Because every everything's in there. So like, oh, this is a video that I made. It's using the transcript. This isn't quite right. Something's wrong over here, or.

This is the book I attached this, the summary's too long. We can summarize that a little bit better so it has direct context without having to look up, find it and bring it back.

Din-Ben Raw: Yeah, I'm, I've been working super hard at trying to build the ethos os the operating system, and, and I've spent countless hours [00:52:00] in chat PT trying to, trying to build this thing and, and I've got 28 years of knowledge poured into this operating system I'm super excited about it. But. I finally got the developer and, and I'm like, Hey, I'm just gonna give you access to this.

You can have everything you want. And, and I'm like, but I'm gonna give you a headstart, right? Like, I got it all in modules and, and, and then I, I'm like, okay, I'm gonna, I got the developer, I told chat, BT you now go ahead and, uh, build out each module for me so I can, you know, and, and I'm like, what about this module?

You totally forget. Like, we've talked about this, we spent hours on this. Where is it? And it. Oh, you're right. You know, I totally forgot. I'm like, oh my gosh. And then all this detail that we went

Ben Tasker Raw: isn.

Din-Ben Raw: it's scary and I'm realizing like I spent countless hours and I've had so much detail into this, and now it's not remembering.

And, and now I have to like prompt, after prompt to try and dig it back out. But I couldn't help but think and, and it's [00:53:00] probably my own fault. What I should have done is. Either created a different project for each module. I, I don't know, or probably not done so much in one chat, like stopped and then created another chat within the project so that I can really like, keep it broken down.

But, but I'm just like floored at what, what happened and I was so angry and I, I couldn't help but think like this, who the ai, this Canvas profile, if I just had all these little modules everywhere. With all of my information now it's all segmented and really detailed in each each module. So I, I, I think I'm liking the concept of this, but I, I, I, I don't

Ben Tasker Raw: I'm visual. I mean, I can see the benefit of that too, right? 'cause it's all laid out in front of you. Where, where the chat thing, right? Like you have to find where you had that information. To your point, you can play around with it, but that takes more time with everything [00:54:00] laid out, you can kind of just rearrange it and, and run it again.

It

Din-Ben Raw: Yeah. Well, there's my plug for Zy if anybody's interested. I think you should, you should go and check it out. Um, so Ben, man, I, I appreciate the conversation. Is there anything else? Um, first of all, is there anything that you are promoting or, I don't know if you, if you do webinars or a book or anything. If you do, I, you know, tell everybody

Ben Tasker Raw: Sure. Thank you Dino. Uh, first off, I hope folks at. Listen today, invest in their personal brand, especially with ai, creating posts, blogs, social media, what, whatever the case may be, has never been easier. You don't even really have to be that good at it or the social media aspect of it. Creating a personal brand has never been easier and ai, it was really important to do that with now because AI is a recommendation engine and of the more posts you have out there, assuming they can be web scraped, the the better recommendation that the AI can [00:55:00] make.

So, so everybody. Think about that a little bit and and how you're gonna use AI to help build your personal brand. Secondly, for myself, if you'd like to get in contact with me to build your own AI application, become more AI enabled your organization. If you need to enable many of your employees, you can find me on ben tasker ai.com, or you can connect with me on LinkedIn, Ben Tasker, thanks for having me now.

Din-Ben Raw: So before I do let you go, um. So let's talk about that. Right? Uh, that's interesting. I didn't actually know that, but like I said, I'm trying to build out this ethos operating system, um, and, and, and I did just hire somebody. But are you for hire for different things like that?

Ben Tasker Raw: Yeah. If an individual wants to learn how to use AI to become more enabled, like, like. Mortgage brokers or even how to set up this simple system. Happy to assist

Din-Ben Raw: Awesome. That's very nice of you. All right, cool. Um, and, uh, they just [00:56:00] find you on all the socials, uh, Ben Tasker, or is there a particular website that, that you have?

Ben Tasker Raw: ben tasker ai.com. There's a booking link. You can book me, you can email me, you can message me through LinkedIn. It, it's all going to the same place.

Din-Ben Raw: Awesome. Well, thank you for joining us. Thank you for taking the time and um, you know, just trying to teach us common folk a little extra about AI and how to make our lives better and our business, you know, stronger. So I appreciate you

Ben Tasker Raw: Thanks for having me, Dino. Yeah, no problem. This is fun.

Din-Ben Raw: and guys, thank you for listening. I really do appreciate it. I dunno if you guys know this or not, but, um, every once in a while, I don't look as often.

Uh. 'cause I just didn't think that it was gonna be that big. But somehow this podcast is totally taken off and, uh, I, I dunno, like a month ago when I looked last, it was top 100 in all of Apple podcasts. So that was just such a huge little. Pat on my back and I'm really proud of that. And I'm, I'm really proud of all the guests that I've had.

'cause I've had some [00:57:00] awesome big name guests on this and, and, you know, and I, and I thank all of you guys, first of all, and all the listeners because you guys are making that possible. So I really appreciate you guys, and if you enjoy what you hear, please do me a favor. Uh, go to the social media, go to YouTube, go to all the places, make sure you like it.

Leave a comment. Um, on Apple Podcast or Spotify or wherever you are, leave a uh, a review. That also helps a lot. So thank you guys. Love you. We'll see you next time.