Are you spending countless hours trying to figure out how to make AI work for your business, while bigger companies seem to have it all figured out? What if you could learn how one of the world's most innovative companies actually implements AI, and discover how to scale those same strategies in your business - without the enterprise budget?
The truth is, most small businesses are drowning in AI tools and possibilities, unsure which strategies actually drive results and which are just burning time and money. They're stuck watching larger companies pull ahead while trying to piece together a working AI strategy.
But today, we're going to change that. I'm thrilled to welcome Michael Todasco, former Director of Innovation at PayPal, where he managed over 100 patents and drove innovation initiatives that transformed how one of fintech's biggest players leverages AI. Now, Michael spends his time helping businesses of all sizes implement practical AI solutions that drive real results. He's going to share the strategies that actually work - no fluff, no complexity, just proven approaches you can implement starting today.
The AI Hat Podcast host Mike Allton asked Mike Todasco about:
✨ Start Small, Think Big: Implement AI strategically by starting with one high-impact area while planning for scalable growth.
✨ Focus on Core Problems: Target AI implementation at solving specific business challenges rather than adopting technology for its own sake.
✨ Measure What Matters: Establish clear metrics and ROI goals before implementing any AI solution to ensure meaningful business impact.
Learn more about Mike Todasco
- Connect with Mike Todasco on LinkedIn
Resources & Brands mentioned in this episode
- PayPal
- James Silberrad Brown Center for Artificial Intelligence at SDSU
- Hanging up on ChatGPT’s Operator
- AI Primer: A Comprehensive Guide
- Explore past episodes of the The AI Hat Podcast podcast
Full Transcript
(lightly edited)
Scale Like PayPal: Simple AI Strategies That Actually Work for Small Business with Mike Todasco
[00:00:00] Mike Todasco: Here's one thing I would say, like, what is the thing where you just got a bunch of data, put a, a bunch of invoices, just put, put your, like literally just drop a file that you have from QuickBooks into an AI and just say, hey, can you make sense of this for me? Am I managing my inventory levels, like be my CFO and just tell me what I can do.
[00:00:20] You could drop in a CSV file or an Excel file or a PDF or what have you, and just start asking it questions like that, and it will go through a similar process and, and you could do this whether you have a finance person at your company or not. Frankly, if you're the finance person at the company, this might be the first thing you should personally do when you have all this data.
[00:00:41] Just start querying the AI and say, hey, what might I be missing here? What trends do you see in this data? AI is great at pattern matching, at being able to connect something over here with something that's over here, and it's just about asking questions again and again and again.
[00:00:59] Mike Allton: Mike's got The Hat, the AI kind. Unlocking power in business, you'll find. Cut through the hype, explore the tech. Every week, new ideas to check.
[00:01:33] Allton leads us on a digital quest. AI secrets, he'll always invest. Tune in tight, don't miss a beat. Knowledge, groove, keep you on your feet. Listen up! To the AI flow. Business growth, here we go. Cuttin' edge insight, that's the plan. With The AI Hat, you'll understand.
[00:02:08] Greetings program. Welcome back to The AI Hat Podcast, where I selfishly use this time to pick the brains of experts at harnessing the power of AI across all areas of business. From streamlining operations and boosting productivity to gaining a competitive edge and driving explosive growth. I. Oh, and you get to learn too.
[00:02:26] Subscribe to be shown how prepare yourself and your brand for this AI revolution and come out ahead. Now, are you spending hours trying to figure out how to make AI work for your business? Well, these bigger companies seem to got it all figured out. What if you could learn how one of the world's most innovative companies actually implements AI and discover how to scale those same strategies in your business?
[00:02:49] Without the enterprise budget. The truth is most small businesses are drowning in AI tools and possibilities, unsure which strategies actually drive results, and which are just burning time and money. They're stuck watching larger companies pull ahead while trying to piece together a working AI strategy.
[00:03:04] But today we're gonna change that. I am thrilled to welcome Michael Todasco. Former Director of Innovation at PayPal, where he managed over 100 patents and drove innovation initiatives that transformed how one of fintech's biggest players leverages ai. Now, Michael spends his time helping businesses of all sizes implement practical AI solutions that drive real results.
[00:03:26] He's gonna share the strategies that actually work, no fluff, no complexity, just proven approaches you can implement starting today. Hey, Michael, welcome to the show.
[00:03:35] Mike Todasco: Well, Mike, thanks for having me. I'm excited to have this conversation.
[00:03:39] Mike Allton: My pleasure. Now, before we dive into these strategies, what do you think is the biggest misconception that you see small businesses have about implementing AI effectively?
[00:03:49] Mike Todasco: I think Mike, you summed it up really well actually in the beginning, the fact that like, oh, this is just for the big companies. 10 years ago that was true, but it is not today. When I was at PayPal, we had an army of machine learning slash AI engineers that were getting paid hundreds of thousands of dollars that were able to do all of these amazing things.
[00:04:10] The thing is, with Generative Ai, with where we're at today. You have access for, 20 bucks a month to the most powerful AI tools that frankly man has ever created. And that is an amazing equalizer. Well, where it used to just be the big companies who had people building these things we're only able to do this.
[00:04:31] You have these tools today and like that is like one of the most wonderful and the most exciting things for small businesses in this wave of AI that we're seeing right now.
[00:04:41] Mike Allton: And what's cool is what, what Sam Altman has said is that the tools today are the worst toll ever be. Yeah. They're only getting bigger, faster, better, cheaper, so on.
[00:04:51] So let, let's talk about PayPal. 'cause when you were there you managed it said over a hundred patents. Yeah. Many of them involving Ai. What would you say are some of the key principles that that guided your AI implementation there? Some of which like would translate to a small business owner.
[00:05:09] Mike Todasco: Oh, good question.
[00:05:10] So I would say, PayPal, we were using AI strategies and machine learning tools because we were forced into it. That's the honest question. The, the honest answer to that because at a, at a company like PayPal, if, if PayPal, especially in the early days, PayPal was ever hacked. And it was constantly trying to be hacked.
[00:05:31] 'cause think about it, it's millions and millions of credit card numbers and credentials and all this other stuff, and there it's a pretty prime target to actually break into it. And you cannot just use traditional computing systems to keep away the bad guys. We built a pretty heavy security and risk AI function over the years to manage sendo way breaks in so that the fraud is under control and all that other kind of stuff. And we did this for a, a, a long time and we learned and we iterated again and again and again. And the one thing I will say is, at least when I was there, I left in 2022. PayPal was never good at using AI and all the other stuff though.
[00:06:13] It was just where it was mission critical is where it found the opportunity. And, and again, if I'm to bring that parallel into the small business world, I would look at whatever it is, the Claude or Chat GPT or Deep Seek, or I don't care what model you're using or talking about. But just ask yourself like, what are the big pain points that I have in my business?
[00:06:36] Is it, dealing with lawyers? Is it, doing planning or something like that? Maybe it's planning schedules or whatever it might be. I would look at my biggest pain points and then look at ways to be able to use AI for that. That's, it's what PayPal did for years is what every other big company does, is what's the biggest, most effective tools at the biggest problems that you have? That's the exact same thing that small business owners need to do.
[00:07:03] Mike Allton: Can you think through one of those examples and walk us through, how you used AI at PayPal specifically and how that might relate to or, or problem that maybe that you solved at PayPal and how that might relate to a small business owner that's maybe going through, obviously not the same problem with Scope, but a similar problem that they could relate to.
[00:07:24] Mike Todasco: I, I, so one of the AI examples that we often gave at PayPal, and this was actually part of kind of the sales script. We, we had this amazing innovation center and we brought customers in there. We got to walk through this and it was it was walking somebody through a transaction. It was done on this big screen.
[00:07:42] You could see the whole map of the world and somebody's trying to buy boots from a retailer in Texas, but they happen to be, they're based in New York, but it's being shipped to California and this is being done from an IP address in Tokyo and all this other kind of stuff. And it would be impossible for any human to be able to like make all those connections and to frankly do that in the amount of time.
[00:08:10] That would be allotted for a, a person to actually say that, yeah, this is a legitimate transaction because a, a lot of what, like your credit card companies, PayPal or Stripe or Square or anything like that, a lot of what they're doing is just making sure that you are you. I. That's why sometimes you put a transaction through, like, why is my credit card not allowing this?
[00:08:29] Well, you're doing something outside the scope of what you normally do, so they think it could be fraud. And so, companies like that use AI strategies to determine, well, what might be going on here? Are there other things in, in that story? And in that story, we used to tell. We would often say like, oh, well your IP address moved.
[00:08:49] We know this IP address was was related to a an airport IP address. And then 12 later you 12 hours later, we see that you add another airport IP address in Tokyo. We know that there's a holiday coming up. You have bought and shipped other things on this holiday, yada, yada, yada. It turns out it's Father's Day.
[00:09:08] You're traveling for business and you're getting something shipped. That's not something a human being would be able to. Pull together and do that in 500 milliseconds or something like that, right? When you would take PayPal valve models. So with that, okay, well, I'm not doing that, Mike. Like my, my business doesn't have that.
[00:09:26] Here, here's one thing I would say. Like, what is the thing where you just got a bunch of data, put a, a bunch of invoices. Just put, put your, like literally just drop a file that you have from QuickBooks into an AI and just say, Hey, can you make sense of this for me? Am I managing my inventory levels, like be my CFO?
[00:09:45] And just tell me what I can do. And you could do this, like all of these models are multimodal now, which means it's not just text. You could drop in a CSV file or an Excel file or a PDF or what have you, and just start asking it questions like that. And it will go through a similar process. And, and you could do this whether you have a finance person at your company or not.
[00:10:07] Frankly, if you're the finance person at the company, this might be the first thing you should personally do. When you have all this data, just start querying the AI and say, Hey, what might I be missing here? What trends do you see in this data? AI is great at pattern matching, at being able to connect something over here with something that's over here, and it's just about asking questions again and again and again.
[00:10:30] And you do that for. 10, 15 minutes with an ai and you might start to uncover things about your own business that you never realized before. And that's where the real value is gonna start to come in. 'cause it's gonna then start to learn this about your business. You're gonna see your business in a different way, and then you could just keep building on that.
[00:10:48] It'll help you better manage your finances, your business, or whatever it is that you're, struggling with at, at any given time.
[00:10:55] Mike Allton: I love that example of what A CFO might do, or any business owner, right, who, maybe they don't have a CFO, but they want the AI to act as a CFO. That's a very core operational thing that, that businesses need to be able to do.
[00:11:09] They need to be able to understand their finances and inventory trends and that sort of thing. Where do you think a business leader should start? What do you think like their first three steps should be if they wanna start implementing AI at that like operational level?
[00:11:22] Mike Todasco: Yeah, so I would say if you're a business leader, first thing is I would use it personally that that is like the, the be all, end all.
[00:11:30] Just get into Chat GBT, get into Claude, get into grok. It doesn't matter. Just get into one of the tools and just start using it. Use it for your own life, make it your default. Use something like Perplexity and make that your default search engine instead of Google search. These are all AI tools. Most of 'em have free versions.
[00:11:50] Just start to realize what they're actually capable of and what they're not. The thing I find out my time and time again is people who don't use AI often overestimate certain things that they're capable of. And, and people who do use it have a little bit just of a better grasp of like, oh, it can do this really well, but it can't do this really well.
[00:12:13] And the only way to figure that out is to play with it. So once you're done playing with it. What I would do is I'd bring into the business if, if you're a small business, say you got like 20, 30 employees buy a corporate license, buy something at the enterprise level that you know all the appropriate employees would do, be able to have, would be able to use.
[00:12:33] And, and, and this is how I would do it if I was really trying to make my company AI first. If I had like a, a weekly meeting, let's just say I had a weekly all hands meeting at my business where everyone gets together and set the agenda and what's worrying them for the week. Say there's 20 people in the room, I would start that meeting with every time, say I'm gonna cold call on somebody and I'm gonna ask you to tell me what's one interesting thing that you did with AI this past week.
[00:13:03] It could be for business, it could be for professional reasons. You already got access. We've given you that, you know you can use it for work. I want you to share one thing about that. Do you want, your team's gonna hate you at first. They're gonna be like, oh God, really? You're gonna make me do this? What's gonna happen after a few weeks, people are gonna start to think about things in a different way.
[00:13:23] They're gonna start to think about it like in an AI first way. They're going to be thinking during the week, oh, what is it that I could do with ai? Maybe I'll try AI to solve this problem for me, and they'll do that. They'll do that again and again and again. And I, I bet you just about anything within a month, you're gonna start to create an AI first S culture.
[00:13:42] 'cause on top of that, what I would do is I'd open up some, if you use Slack or teams or whatever it might be to share messages, I would put up an AI channel and just say, Hey, people, share what you're doing in this AI channel. Because that's how we learn. We learn from each other. The, the difficult thing about these models is they have infinite capabilities and possibilities of what you can do with them.
[00:14:05] So it immediately freezes people up. We don't, we, because we can do anything and everything. We do nothing. That is a very natural human thing. You got to push your team, push your company past that so they see, oh, there's all this cool stuff that I can actually do. Get them past that, that that stopping point, that little barrier that they have.
[00:14:27] And then you're gonna actually start to see, well, we're building a company that's using this for a whole bunch of different problems that they weren't before.
[00:14:36] Mike Allton: I love that so much 'cause the underlying truth that, that, that's behind what you were just talking about is that you need to have a leader who's embracing ai Yes.
[00:14:46] And empowering their team. Right. Who's saying, you know what, okay, we are going to get Google Workspace with ai, or we're gonna get a ChatGPT Teams account. You're all gonna have access and you're all gonna share at the weekly hall hands, because I run the agenda and I'm gonna carve out that time and I'm gonna make sure that's there.
[00:15:01] And I've got the keys to the Slack channel, we're gonna create a special channel. I I think that's fantastic and I couldn't agree more. Do you think that creating some kind of an AI policy document should be part of that initial strategy?
[00:15:15] Mike Todasco: Yeah, I think it's important and, and I don't know how much it depends on how sensitive the data in your corporation is.
[00:15:21] Yeah. I think generally speaking, there should be some level of parameters. Just like I am sure within your company there's something that says, Hey, you shouldn't be using your personal email to email customers or, or email vendors or anything like that. You use your corporate email that's, that's, within our servers and all that.
[00:15:41] You would never use your personal email for that. It's probably a similar set of rules. Like, you shouldn't share this, you should, this is okay to do. Just give some level of guidelines. And here, here's the great thing with that, Mike. If you don't know what, how to start with those guidelines, like literally just ask the AI.
[00:15:57] Just say, Hey, this is my business, this is what I'm concerned about. Can you give me bullet points? That would be guidelines that I could create, that I could send to my employees that would help them have an AI policy for our company. And boom, it's gonna generate that in 10, 15 seconds, you look through that and say, ah, I gotta tweak this, tweak this.
[00:16:15] And I bet you within five minutes you have those guidelines that you could then use. Like, look, if you're at a place that as a lawyer or a privacy team, great, but I'm assuming most people don't have that. So just let the AI do it for you.
[00:16:27] Mike Allton: Yeah, I was just teaching a, a workshop, in fact, where I was, I was showing them some custom instructions and I was mm-hmm.
[00:16:32] Showing them the files that I had attached to that. And I had attached a target audience document and a, and a brand voice document. And they're like, well, oh, we don't have a brand voice document. How do I create that. I was like, well, let's just start with AI and you can have AI walk you through how to create it, and then you can ask the AI to turn that conversation into the custom instructions that you need to provide the AI for everything else that you wanna do that uses your brand voice.
[00:16:58] So that's,
[00:16:59] Mike Todasco: And, and, and I'll say what you, you said a, a great thing. One of my favorite hacks with chat GPT specifically is to use their advanced voice mode. And this is just a very natural, very conversational thing. You can have, you could actually say to chat GPT Advanced Voice Mode, Hey, here's my company.
[00:17:17] I want you to actually interview me about this. Mm-hmm. And just ask me a bunch of questions for that. And you could talk to it just in your own natural language without even typing. And then at the end of it you say, okay, if you're done with questions, can you actually summarize this in something that I can make into a document?
[00:17:34] And then it'll just take what it, the interview that it did over the last 20 minutes or whatever and put that into a real document. For somebody who, like, I've started writing books that way. I use the, the audio features, we use it with my kids in the car when we're driving to school to quiz them on a, on a test that they have coming up.
[00:17:53] Any of those kind of things like that is a killer, killer feature that people do not use enough. I.
[00:17:59] Mike Allton: Love that use case folks, write that down, practice it, do it yourselves. But one of the things that I know you've mentioned online, Mike, that I've seen is you've, you've said that AI should be viewed as augmented intelligence rather than artificial intelligence.
[00:18:13] Can you walk us through what you meant by that and why that distinction matters for businesses?
[00:18:18] Mike Todasco: I hate the name Artificial intelligence. I actually wrote a whole piece about how bad a decision it was when they gave the name back in the 1950s at Dartmouth of artificial intelligence. There's actually a record of all the other things that they could have called it and, but like this was the most sci-fi and was a way to get attention.
[00:18:37] We were in the middle of. A race to space with the Soviets and to do things that were very futuristic and sci-fi that generated a lot of interest. So that's why one of the main reasons why we call it artificial intelligence to this day. But here's the thing, like there's a lot of negative connotations that come with that.
[00:18:55] It's not its own being, it's not its own species. It's nothing like that. The artificial intelligence that exists today, it's really just augmenting us and making us better. I write a lot and I use AI all the time in helping me do research, helping to proofread what I'm doing helping to give me feedback and things that I'm writing and so forth, all that kind of stuff.
[00:19:22] I also just at, at times I will have an AI and I'll just say, Hey, can you write a book for me? And I'll tell you when you just do that. It's awful. Like and I publish these books in fact, so if anybody wants to go to Amazon, you can go to look up Alex Irons. That is my pen name that I use for all these AI books.
[00:19:42] I give all the money to charity so I'm not, so you could buy 'em, they're not very good. But for 99 cents you could buy one of these things and see from 2022 until today, what it looks like over time when AI writes a book and look, they're not very good.
[00:19:57] Mike Allton: Please tell me. These are like combat war novels.
[00:20:01] What did these
[00:20:02] Mike Todasco: Oh, so I will give it a, a topic. So it is varied. Everything from I did one about like all 50 states and unique stories from the 50 states. I did when Mickey Mouse went into public domain, I had to write a Mickey Mouse story for me. That was within the public domain guidelines.
[00:20:18] It's a whole bunch of different stuff in there. And then, yeah, not too much about AI taking over us or anything like that.
[00:20:26] Mike Allton: The pen named Alex Irons struck me as someone who's gonna write about, green Berets and, seal missions and stuff. I don't know. It struck me well,
[00:20:34] Mike Todasco: well, it's funny.
[00:20:35] So the AI actually gave me the name Alex Irons. I was like, I. I want, because I think I asked it. I want names that have AI in like the first, first name, last name. Got it. So I want the initial to be ai and then also I wanted to, the.com is available, so I think if you go to alex irons.com, it'll forward you to the the the Amazon page.
[00:20:56] But again, I'm not making money. They're no good. But if you wanna really look at, it's, you'll, you'll help feed some people if you, if you happen to buy one of those on Amazon. Where, where were we? What was the.
[00:21:08] I lost track, Mike.
[00:21:10] Mike Allton: No, that, that was terrific because you were pushing back on the idea. I think that a, a lot of folks have that using AI is, is cheating. Ah, ah. That they shouldn't be using it. And, and I, I, I think that was just great. That, that you, you, you pushed back on that folks were talking with.
[00:21:27] It's a tool. It's a tool. It's tool. Unlike
[00:21:28] Mike Todasco: other tools. That's right.
[00:21:29] Mike Allton: Absolutely folks, we're talking with Mike Todasco about the lessons and ideas businesses of all kinds can take from PayPal and similar enterprise initiatives. And I've got seven more questions you're not gonna miss, but first, a quick word from our sponsor.
[00:21:41] This episode of AI for Business Leaders, The AI Hat Podcast is brought to you by Magai. Tired of juggling multiple AI subscriptions and trying to figure out which tool to use and when? Magai simplifies AI for businesses by offering access to all the leading language models and image and video generators for one low price.
[00:22:03] With Magai, you can create custom instructions to personalize AI for every business need, and use them across all LLMs, including ChatGPT, Gemini, Claude, and even Deep Seek. No more building custom GPTs or gems that are limited to just one platform. Magai also lets you create customized workspaces for your entire team, giving everyone in your organization access to the same AI assistance and tools and prompts ready to streamline your AI workflow and empower your team.
[00:22:31] Tap the link in the show notes to visit Magai today to learn more, and now back to the rest of today's episode.
[00:22:40] Have you ever heard of Magai Mike?
[00:22:42] Mike Todasco: I, I have nots, but I, I feel I need to go to the link in the show notes and check it out.
[00:22:47] Mike Allton: Well, you gotta check it out. I, it just, I just occurred to me to ask because a friend of mine, Dustin Stout, is, is the founder and he's not far from you.
[00:22:54] He's just a little bit further the south in California from you. But I was also thinking it, I'm, I'm way overdoing redoing that video. 'cause I don't talk about the thing that I like the most about Magai, which is the fact that it's a wrapper for all the other LLMs. So using Magai. You can access Chet GPT, Gemini, Claude, Llama.
[00:23:15] Now you can access deep seek. He's got 'em all in there at the paid level, but you only have to pay Magai one $20 month fee rather than paying all these other ones. Right. And you still get the custom instructions and everything. So it's a super fantastic tool. That's great. Definitely check it out. But yeah, it, it, it answers that, that probing question that a lot of businesses have is, well, what, which one am I supposed to use? And we've touched on it already. You said, just use one of 'em, but still, which one? But I think. Where I'd love for you to go with next is if you could talk about what you think some of the biggest mistakes businesses are making when they're trying to implement AI beyond just what we've already talked about today.
[00:23:54] Now they want to do actual, like automation. They want to do some workflows and that sort of thing. What are some of the mistakes that you see them making?
[00:24:03] Mike Todasco: I think the first thing that I often see businesses making and that's a mistake, is over optimizing. I, I think that's in reality, if you're going from zero to some sort of AI model that you're going to be using, like it almost doesn't matter which one you're going to be using.
[00:24:20] They're all going to be extremely good for whatever it is that you're trying to accomplish. And I think a lot of times businesses will get frozen. They will, say like, well, I don't know which one of these to use or anything like that. So they just don't choose any of 'em. Or they could say, well, I think it's just gonna get better in a couple months.
[00:24:38] Like, I heard this one thing is gonna be coming out. There's some rumor, like, why don't we just wait for that? We'll wait for it to see it come out and, and see what happens then like. There's always gonna be an excuse not to implement these, that that's just the reality of what we're we're facing and like with these cycles and how quickly things are evolving.
[00:24:58] I. So the biggest thing that businesses can do is just use it like literally just jump in. You pick one of the models, do a little bit of research on it, just see like, okay, we're going with where? With Anthropics Claude. We're going to OpenAI ChatGPT, we're going to open source a Meta Llama model. We're going to use Magai.
[00:25:19] We're going to test them all. We're going to do whatever it might be. Just just pick on something. You could revisit it again down the line, but the productivity, the benefits that you're gonna get for just starting with something and one of the major models out there is going to more than offset any time you're gonna just send, spend frozen.
[00:25:39] Just like, ah, I'm trying to find the perfect model. Just find one that's good enough.
[00:25:45] Mike Allton: Love it. Fantastic advice. And I know you're sharing that kind of advice through San Diego State University to, to students and I'm sure businesses of all kinds. I, I'm sure you've heard lots of success stories of businesses that have, have effectively scaled some Ai.
[00:25:59] I wondering if you could think of one that you could share with the audience.
[00:26:03] Mike Todasco: Well, look, I, I think the reality is like, most Ai we, we've been working with AI except it was not out in front. AI's always been in the background or, or at least for the last 10, 15 years. So a company like PayPal, it was making risk decisions.
[00:26:20] If you're using any social network out there, the AI and machine learning is going to be generating your customized newsfeed. All of these things are using AI as part of that, and today we have, these wonderful tools of Generative AI that are definitely put out front to people and that people can start using them and so forth.
[00:26:45] And honestly, Mike like. I don't think we've even begun to see what the benefits of this truly are. I am, I'm hearing stuff out there and, and actually talking to people at, medium or larger sized companies that seem to be doing a lot with these tools as far as, more effectively using 'em for customer service. I think Google said that 20% of that code that engineers were writing is now being done by ai. I might begin that number a little bit off, but like that was like the last number that they put out there. So, and, and this is like the same kind of things that you could buy if you have engineers in your firm, you need to be using these tools.
[00:27:29] Yeah, I don't think there's anything I. Any kind of background, any kind of function that you have within a business, as long as you're working with a computer that you're not going to have significant benefits in doing ai. Again, law firm, legal, things like that, where I see huge benefits for businesses just saying, Hey, we're gonna do our first draft of contracts, and maybe some of these that are more sensitive.
[00:27:59] We'll do a first draft and then send to a lawyer, or whatever it might be. And or, or even just interpreting contact tracks is a huge thing. So coming in the other direction. Yeah. Instead of having to say, Hey, can you read through this whole thing? One of my favorite things to do if I get something, I'm just say, I throw it into an AI and say, Hey, is anything out of the usual out of the ordinary here?
[00:28:18] Anything look weird in this contract, non-traditional that I should watch out for? And if you think about the AI has seen thousands and millions of contracts. So it knows if something's trying to be snuck in there people have done, it's, it's called a needle in a haystack test. Mm. So you can actually take like a book that's in public domain, say like you take Huckleberry Finn, insert, on page 232 A, a line about pizza Hut is having a sale. And or buy one get one free, and then put it into Claude. And people have done this before. And then say, Hey, is there anything weird in this book? And it will almost always come back to you and say. Yeah, on page 232 there's an ad for Pizza Hut and it does that within seconds.
[00:29:06] Yeah. Like it knows how to do that stuff. It is incredible at finding that, so, like, again, it depends on your business use case, but like for me for me being a non-lawyer, the legal stuff is where I would turn to especially contract interpretation.
[00:29:23] Mike Allton: That is, that is great advice, and you and I both know that's going to result in save time.
[00:29:28] That's gonna result in potentially finding issues before they become issues. And may, maybe somebody's got a clause that they've put into the contract that really doesn't gel with how you do business, and that's gonna help you find that right away. But I'm wondering, on behalf of business owners, how do they actually measure for themselves these benefits?
[00:29:48] How do they measure the ROI of implementing ai?
[00:29:51] Mike Todasco: Okay, so I learned a lot about, so I, my background is in finance. That's where I started my career. So I was the ROI guy for, for many a years. And then I, I moved into, one of my jobs at PayPal was running innovation there. And of course, you know that in the back of my mind I'm thinking, well, what about, what's the ROI?
[00:30:09] How is that going to actually matter? And the way I found around that, and there, there was a way around it, it made a lot of sense, at least for the situation I was in at PayPal. Was our innovation team and our innovation costs were so tiny, infinitesimally tiny, that it wasn't a worry, it wasn't a concern.
[00:30:29] Hmm. It wa it was just a simple rounding error with something like that. So everyone left us alone. They knew we were like, it was like sometimes just me, sometimes me, and one or two people or whatever. But we kept it tight. We kept it lean and that was part of what we wanted to spur and, and inspire in the company.
[00:30:45] Like, Hey, you don't need a whole bunch of money to be thrown at you. So if I'm a small business owner and I'm thinking about AI and these tools. All I would do to start. Is, I would say, do you want, we're not even gonna worry about ROI for six months, for 12 months, we are going to spend 20 bucks a month, and that's generally as much as this costs to get Chat GBT, Claude, whatever it is.
[00:31:11] Again, the, the tool doesn't actually matter, just like the fact that you're using it matters. And we're going to revisit that after six months. So keep the cost low in the beginning. The last thing you wanna do is say like, well, we're gonna do the $200 a month plan. Yeah. And we're gonna do this, this, and that.
[00:31:27] Like, no, no, no, no. 20 bucks a month that is gonna get you everything you are going to need to get started six months from now. I think it's worthwhile to take a look back and see how have your workflows changed? Are you using less? Legal costs? Are you giving better service to your customers?
[00:31:45] Are your employees able to work less overtime? Whatever it might be. Like that's important to you for that. But give it six months to do that because it's going to take a while. And if you are finding after six months, do you want, I think this AI is adding no value and like no one's actually using it.
[00:32:02] And despite what Mike told me in the beginning of, Hey, let's make this an AI first company, then do you know what? And I guess you probably shouldn't do it. I think that's the honest answer to that. I think more likely what you're going to find out is that people are going to be using it and you're going to then be able to make a decision.
[00:32:22] Okay. Should we do a couple other things like, it works real well here, but man, it would be amazing if we had agents and could do this and so forth. But honestly, I would not get into any of that kind of stuff until at least like the six month mark of just working with the traditional systems that are already super low cost.
[00:32:41] If you can afford, 20 bucks per head you should be able to afford that no matter what your business is. And again, look that as an investment. And if a few months that there isn't value, then you can reconsider it.
[00:32:55] Mike Allton: That is such smart advice. I appreciate you sharing that. And I wonder if you could just for our last question, just kinda look ahead and say what AI emerging trends you're seeing, like you've talked about agents. We haven't really talked about operating, and I'll, I'll preface this by saying on this particular show, we don't often get into the latest and greatest things, and we avoid as much as we can. But that said, it, it is unavoidable to talk about, specific AI models and that sort of thing.
[00:33:24] But again, just, what's coming do you think that small business leaders in particular should be thinking about.
[00:33:31] Mike Todasco: So, so the last week, I, I will touch on the two things that you mentioned. So, yeah, so Deep Seek and Operator came out. Deep Seek is a Chinese model that had been launched and it is extremely low cost.
[00:33:44] It is also Chinese model, so there's a lot of censorship and so forth that it is part of that. So, like, realize that, don't be surprised when you see that. But it is much lower cost. I think Deep Seek is not as special as frankly, the methods that were used to make DSE are going to lead to much lower costs for these models going forward, and frankly, much higher performance.
[00:34:05] And that's the most exciting takeaway for most folks. Operator is open AI's first chat bot. And so I'm actually writing a piece that I'm gonna release in the next day or so. It should be live by the time folks are watching this about like my first week using operator and I will say, and so operator, you have to spend the 200 bucks a month for the pro subscription with chat GPT.
[00:34:28] I've been using it and I am so disappointed. I couldn't tell you how disappointed I am with it. Like I've used other agents stuff. It's pretty useless. It's actually shocking to me that they release this, I was struggling with it and I said, Hey, operator, what should I actually do with you? And I literally was, and I was like, oh, I'm really good with travel.
[00:34:47] And I'm like, cool. Tell me like in June of family four, we're gonna wanna travel somewhere to Europe. Give me some prices like for direct flight from San Francisco to Europe. Something like that is what I gave it. It went back in a thought for 24 minutes. I'm not making this up like this is in my article.
[00:35:05] It went back, thought for 24 minutes. Again, no problem. AI goes to work for 24 minutes and it comes back to me and says, I couldn't find any flights for you going from San Francisco to Europe in June. I. Here's a flight going from Orlando to San Francisco over the timeline that you required. And I was like, wow.
[00:35:23] I swear to God. Like that was, and this is after other frustrating stuff. I'm like, are you kidding me? It took 24 minutes to give me like a totally useless thing like. It's half baked. To, to my friends at Open ai, I just ask like, what are you doing right now? Right? Like this, this should not have been released, at least at the $200 tier.
[00:35:43] Maybe it could have been free for people to test it, but like, oh man. But that's it. Where are we going? Again, to your original question, Mike. These things are gonna work. Operator isn't gonna be always this bad. There's going to be other agent type software that will be able to spend 25 minutes and actually give me a killer itinerary and saying, Hey, we found this one flight and we also found this flight, and it uses miles to do that and we know you have a bunch of United miles or whatever it might be.
[00:36:10] And doing that and that and that. That's the kind of stuff that's exciting. So I think. We are gonna see cheaper models. Deep seek is an example of that. We are going to see more agents in our work lives and our personal lives where these things can run semi autonomous autonomously in the background and they could like work to some sort of task that is emerging today.
[00:36:35] And we will start to see more and more of that in the months to come. I, I think 2025 is going to be just immensely exciting in the world of AI and it's only gonna accelerate more and more. I think that that is the reality. Like, and, and that's, may, may we all live in in exciting times.
[00:36:53] Mike Allton: Right?
[00:36:55] Yeah. I couldn't agree more. I was giving my client the same advice earlier today. 'cause we were talking about operator and we were exploring some potential use cases, like could they use operator to do a LinkedIn sales navigator search and start to surface prospects and then enroll prospects in a workflow that uses a different app and so on.
[00:37:14] And I was like, theoretically, yes. It should be able to do that, but in practice, probably not today, probably in three to six months. Yeah. So.
[00:37:26] Mike Todasco: Yeah, it's the right integration. It's the right whatever. It'll get there. It'll, so you can dream these things. You might wanna just test to see, this is like the problem of like, all the characters and DALL-E one and DALL-E two having, six fingers and things like that, and just the form faces and all that like.
[00:37:43] It was obvious to anybody who is seeing the progression, like that's not gonna last forever. And now, frankly, most images you have don't have six fingers anymore. They've figured out all of those problems. There's other things. Images are getting better and better. It's the exact same thing with the agents like operator.
[00:37:59] It's the exact same thing with all these models. What is true today? And any of the shortcomings is very likely not going to be true tomorrow. And the only way to find out is frankly, just for you to use these models and just to keep testing and seeing what they're capable of.
[00:38:16] Mike Allton: Could not agree more. Mike, you have been fantastic.
[00:38:19] This has been such an interesting deep dive and, and co to what PayPal had been done and your lessons from PayPal, which is. Fantastic. So I appreciate you sharing that with us. For folks who wanna learn more about you or how you could potentially help them, where should they go?
[00:38:35] Mike Todasco: Yeah, I think just my LinkedIn is the best place.
[00:38:37] Just Mike Todasco on LinkedIn, you could find me there. I have a little newsletter where I try and explain AI stuff to, to common folk like myself. And so if you're into that, you could sign up. It's all free and everything. I'm there and happy to answer any questions that you have. If people wanna just DM me there as well.
[00:38:56] Mike Allton: Terrific on behalf of the common folk. Thank you, Mike. We will have the links to everything we talked about in today's show notes, as we always do. That's all the time we've got for today, friends. But remember, if this is your first foray in the world of AI in business, I've got the perfect resource for you in the show notes below the AI Primer designed to quickly get you up to speed on the terms, the tactics, and the underlying technology that's powering this AI revolution.
[00:39:19] Check it out and I'll see you in our next episode. Welcome to the grid.
[00:39:23] That's a wrap on another episode of The AI Hat Podcast. I hope you've gained some valuable insights and actionable strategies to boost your business with ai. If you enjoyed this show, please subscribe, leave a review, and share it with your network. And don't forget to visit the ai hat.com for more resources and to connect with me.
[00:39:42] Until next time, keep innovating and remember, the future of business is human and it's powered by AI.
