Auto-generated transcript
Scale Smart, Grow Fast: 65. Future-Proof with AI: PayPal's Innovation Leader on Scaling Smarter
This is an auto-generated transcript, made with speech recognition and lightly cleaned up for readability (filler words and false starts removed). Speaker labels were assigned automatically, so a brief interjection can occasionally land inside the other speaker's turn. The original recording is the authoritative version.
Mike Todasco 0:00
This is that world times 10 because what we're creating with these AI, we don't even know what they're capable of. If we just stopped working on them today, we're going to have years and years of growth just as we figure out what the models today are capable of.
Harley Green 0:15
Welcome to Scale Smart, Grow Fast, the podcast designed to help high-performing entrepreneurs and executives scale their businesses smarter, not harder. I'm your host, Harley Green, founder of WorkerGenix, and every week we bring you strategies to leverage your time, team, and technology so you can grow faster, increase profitability, and free yourself from the day-to-day grind. Through expert insights, real-world success stories, and battle-tested frameworks, we'll show you how to delegate effectively, systemize your business, and unlock the power of ultimate executive assistance. If you're ready to scale your business, work less, and achieve more, you're in the right place. Welcome back to the Scale Smart, Grow Fast podcast.
We have a special guest today, Mike Todasco, and he's going to share some amazing tips about AI. And I'm sure as many of you know, AI isn't just a buzzword. It's reshaping how businesses operate, scale, and make decisions. In this episode, Mike Todasco, former Senior Director of Innovation at PayPal and visiting fellow at SDSU's AI Research Center, shares how entrepreneurs can harness AI to optimize efficiency, enhance decision-making, and stay ahead in a competitive market. If you want to future-proof your business and lead AI-driven innovation, this is a conversation you won't want to miss. Mike, welcome to the podcast. How are you? Harley, I'm doing great.
Thanks for having me. Well, thanks for being on. Can you maybe share with our audience a little bit about your background and how you came to being an expert in AI?
Mike Todasco 1:39
I started my life as an accountant way back in the day. I was just I didn't know what I wanted to do out of college. I knew I liked numbers. That's about all I knew. So I went into like finance and accounting. I did that for years. I was always into technology. I always like to tell people like I actually had a robotic lawnmower like 22 years ago that I was working on and like it was by this company Friendly Robotics. I was always the weird guy who had all these like tech toys and all that, but I was, you know, an accountant by night by day and doing that stuff by night.
At some point in my career, started a business. It was an online e-commerce marketplace. It failed, but that led me to PayPal where I was at for 11 years. And at PayPal in the 2010s when I was there, we were doing a lot in the world of machine learning or artificial intelligence and I just got plugged into very much that world. Eventually, I got the opportunity to run innovation globally for PayPal, which was super fun. I just got to work on like new and emerging tech all the time. And then when I left PayPal 2022, I was already playing a lot with like GPT-3 and the tools that were out there.
Midjourney just got released, which is an image generation tool that summer. And I was just like, my god, this is as a non-engineer, it was really clear to me that this was making the 99% of people who are not engineers on this planet, it was giving them superpowers that like we never had before. Because I would see what the machine learning engineers who were getting paid 600 grand a year could do at PayPal, but for normies like me, there was never a way to do that until GPT-3. And so I just started writing about it and I've been doing that ever since and just trying to spread the word that like, hey world, like it's changing now.
And I got involved with San Diego State, so I do a lot with the university there, a lot of teaching and things like that in that world as well.
Harley Green 3:27
Now, our audience of business leaders have probably been thinking about AI and many of them maybe have adopted a little bit. They've played around with ChatGPT or some of them maybe are just like, ah, I don't want to, you know, mess with this. It's new technology. I'm kind of scared of it. So with your experience in AI innovation, how do you see AI transforming business operations, especially when it comes to that executive support?
Mike Todasco 3:47
Let me start with a story. When I started college, it was back in '95 and I just happened to be going to University of Illinois in Champaign. And when I got there, one of the lucky things was like everything was online there. And I had literally never been on the internet before I got to college. And it was one of those things I was like, and when I came back and told friends about it, they're like, what are you talking about? Like, why would I use this? It just none of it made sense to like my high school friends or anything like that. But it was clear to me like, oh my god, I'm like living in the future.
And it was just happenstance that that happened to be. This is that world times 10 because what we're creating with these AI chatbots, we don't even know what they're capable of. If we just stopped working on them today, we're going to have years and years of growth just as we figure out what the models today are capable of, let alone the next version beyond that. So from a standpoint of like executive support and things like that, like it's incredible what these things can do. I am constantly just having conversations when I go for a walk with an AI, brainstorming, thinking of better ideas, having it take notes for me, action items, to-dos, all this like and I'm in the car, I'm wherever.
And this is all available for like 20 bucks a month with a ChatGPT or Claude or Gemini subscription. It's not even expensive. It's ridiculously cheap. I always tell people if the tools in 2025 existed back when I was at PayPal like in 2021, we'd probably pay $100,000 a seat for these tools. And now they're available like 20 bucks per month per person today. The one thing I just tell people is use them. Like it doesn't work if you just dabble because you're going to find out you'll find shortcomings and they have tons of shortcomings. They're not perfect at all. And that's one of the things you got to figure out where they're good, where they're not good.
But like just start to integrate it into your workflow, into your life, into all these different things and you will be so far ahead of the you too will be living in the future while your competitors and everyone else is still living in, you know, 2021 before this was widespread. Yeah, I love how
Harley Green 5:50
You said just start using it. And what are the biggest challenges that companies face when integrating and trying to have that just start using it these AI-driven solutions and how can they overcome those? So
Mike Todasco 6:02
I think the first challenge is a legitimate one is like is data leakage and other worries about security and privacy and all those kind of things like that. So what companies need to do is they just need to buy a subscription to these services. They need to sign up with Copilot or Claude or ChatGPT. It frankly, it doesn't even matter. They're all good. They're all terrific. You need to have a tool that your employees can use and start using in a way that's not going to be compromising your data and information. Most companies out there are already using cloud services by these big tech companies.
So you're sharing data in some way for that. So like this isn't really that different than that. It gets into some nuance of that. But for the most part, you shouldn't worry about that. Do it in a proper way. Give your employees a way to do this because they'll be using this even if you don't if you say they shouldn't. There's been some statistics about the number of people in companies that are using ChatGPT on the side where the company doesn't allow it. And it's something like 60-70%. Allow it in a proper way and then encourage your employees to share that. Like if I wanted to be an AI-first company, one of the things that I would do is say we have like a monthly all-hands or something like that.
I would pick out two or three people from wherever in the company at random to just share how they're using these AI tools that I have for my company on a monthly basis. That is just one way to kind of show, hey, we need to be AI-first. Because here's the thing, like while the tools are incredible today, in the next 18 months, they're going to be so much better than they are today. But you need to start using them now. Start to get familiar with them. Start to build that into your workflows and so forth. And just start to see how it could help you.
I'd love to say, well, you should do this, and this. But like you kind of got to figure that out yourself. And ultimately, you need to just create a culture and environment where employees are just going to be allowed to freely play and try these things and see what works and then share what does.
Harley Green 7:57
Yeah, I love that you mentioned that. Something we've actually done in WorkerGenix is give training to our employees. And then during our like weekly or bi-weekly kind of social get-togethers, we open it up and let people share how they're using AI that's new, whether it's, you know, the latest video editing tools or how to, you know, have ChatGPT do this new function for you. And it's really exploded the opportunity for everyone just getting to understand it because it's such a new technology. People sometimes need some prompts to guide them to leverage it.
Mike Todasco 8:27
It's a paradox. When we have infinite choice in front of us, we freeze. On Stanford up the street from where I live here in Silicon Valley, one of the first studies on this was done with Jelly. And like literally at a grocery store, if you only had, you know, two or three jellies on the shelf versus like 30 options, people actually buy more when there's just two or three options because people freeze when there's too much choice. The problem with AI is there's literally infinite choice. There's no limit to the amount of things that you can do in this. And so people's natural reaction to that versus what we've dealt with software in the past where it's binary where I mean literally there's there's a happy path to go down and it's based on ones and zeros.
This is based on mathematical probabilities that have no real end. So like people just freeze up. And that's why we need to see examples from other people. Examples where, oh my gosh, this person in accounting used this software to vibe code, a thing where like now they could just scan invoices and it's going to take that information to put it in there and it's all done on that person's computer. And like, well, then what could I code with this? Like there's so many possibilities. And we need to see examples for us as humans to be able to then make the connection and say, oh, okay, that's what it's good for or that's an example. And again, that only way that comes from doing that is sharing.
Harley Green 9:47
You mentioned the decision making. What are some practical ways AI can be leveraged to help leaders streamline operations in decision making? So
Mike Todasco 9:56
One of my favorite and I don't know if this is actually streamlining, but frankly, one of my favorite things to use AI is to be the devil's advocate. If I'm putting out an email, if I'm doing whatever in an organization, you know, maybe I'm making a strategy jock or like this is a five-year strategic plan and I'm a leader and I'm working on this, I love just feeding those documents into an AI and say, hey, be my devil's advocate. What are the holes here? What are the alternate scenarios that I'm missing in this future? What are the gaps here? Once a strategy is set to, you know, to kind of get everyone in line and stack up, but while it's still being developed, that is the time where you need to avoid groupthink, where you need to poke holes in that.
And in some organizations, that can be hard. So that's why if AI isn't worried about is my boss going to be mad or any of these things that we as human beings like, oh, well, maybe I'm just going to go with the group and say, yeah, that's a great idea. No, an AI can give you a much and especially if you give it direction. And that's one of the things when you're prompting an AI, say, hey, I want you to disagree. I want you to find the holes in my thinking here. That will just make it stronger if you really have something poking holes in that to build out that strategy or whatever you might have. And that's one thing that I have used and I have recommended a lot because AI can be extremely good at doing that.
Harley Green 11:14
Yeah, that's a good point because a lot of times if you're trying to do a brainstorming session, especially if you're a leader and you've got subordinates working with you, they might be hesitant to maybe contradict what the boss just said versus if you're doing that with an AI session, they don't care about that. They're not worried about getting fired or promotion or political issues like that. Yeah,
Mike Todasco 11:33
No, that's exactly right, Harley. I mean, so that is the great thing. And that's why people, as we said in the beginning, people need to play with these tools to know what they're good for, to know what they're not good for. And that's an example where like, yeah, it doesn't care about any of that other stuff that humans care about. So therefore, like let's use those that lack of empathy, that lack of whatever you might call it that it has to actually help us do what we want, which is to make our business and our strategy better and so
Harley Green 12:00
Forth. So Mike, what are some of the most exciting AI advancements happening right now that business owners should be paying attention to? Everything.
Mike Todasco 12:07
Like that's the short answer, but let me give you a specific example. I was actually teaching two weeks ago and I had a Monday and a Wednesday session in that class. And Monday night, I went and gave a presentation about two hours long and all that. Tuesday, ChatGPT announces OpenAI who makes ChatGPT announces a new image generation that they have. And I had a whole section on image generation, what it's good at, what it's not good at, blah. I had for the Wednesday class, I had to rewrite that entire part of my presentation. So the exact same class, two different sessions, the world is changing that fast.
And I told them and I told the Monday class as well, like, hey, by the way, like what I told you yesterday is now wrong. It needs to be updated. And that's just kind of the reality of this. So what I'd say to people is this, like don't get too overwhelmed with what's the latest and the greatest. Just get into the tools and just use them. Gemini, ChatGPT, Claude, they're great. Like I pay 20 bucks a month for all of them. Just pick one and go with it. Don't worry like, but the score is a little bit better on this one and the score like, you know what, it's going to be fine.
Like there's so much competition in this market, prices are so low. Just kind of latch on to one and use it in whatever way that you see your company. There's a lot of other things that are happening. So like that's the foundation model. A lot of companies are building other services around these models where there's a lot of excitement. All companies like this, if you're an engineer out there and you're not using one of those tools which tie into the foundation models I said before, like you're probably being left behind. And what's happening today for engineers, for software engineers is going to be happening to a lot of other business functions in the very near future.
If you're in finance, if you're in legal, if you're in wherever, there are already tools out there. They're good. They're probably not as good as the ones the engineers have, but in 12 to 18 months, they will be. So just kind of keep that in mind. Dip a toe into the other like area specific tools that might help you in your job. I want to go much further than that. Just, you know, one foundation model to use for everything and then maybe one domain specific one that's going to help you and your team do whatever it is you need to do.
Harley Green 14:20
I love that advice is just kind of keeping it simple while still using an advanced tool. You kind of mentioned how things are changing so quickly. What can companies or leaders do in this market where it requires constant innovation to fully leverage AI and maybe take advantage of having like executive support with a virtual assistant that knows AI to stay ahead?
Mike Todasco 14:39
Look, if you are a VA today and not using AI, just as I said with the engineers, you're completely being left behind. So if I were a leader and working with a virtual assistant, I would actually test them on their AI abilities. Like I would just kind of say like, hey, what is it that you can do? What is it you can create? If I'm trying to hire between two VAs, I'm going to say, hey, I'm going to pay both of you for four hours of work. I need you to do some superhuman type task. I need you to take whatever it might be, a book I wrote previously, and I need you to work with an AI to rewrite that for 2025, whatever it might be, and see what they come back with.
And like give them a day to do that or whatever it might be and just see what they can do. A great virtual assistant would probably use a tool like NotebookLM and they would have like a different instance of NotebookLM. So NotebookLM for people don't know, it's made by Google. It's using a slightly different technology called RAG, retrieval augmented generation, where it really keeps core to the source documents that you feed into it. And so what that means is like you can have a different setup for each person that you're supporting as a VA. You could put every call in there. You can use a tool like Otter or Fireflies to transcribe every call.
Put that all into there. You have this whole corpus of information that you know about the person. And that will just have you better answer questions and queries and, you know, whatever it is that you might be doing, whether it's helping them book travel, helping them summarize action points, helping them build a deck, whatever it is, you could kind of put that in there, identify a person's style, and just be so much more effective with it. So yes, if I'm hiring a VA, I am 100% looking for how they're using AI. And I don't want them to hide it. Like my expectation would be that they are using AI.
And if you're a VA, I would say be up front with that. Say like, yeah, I'm using AI. And you know what? I use it better than any of the other VAs out there. And I could do three times the work in the same amount of time as any of these other ones. And here's why. That would be the pitch that I would give.
Harley Green 16:45
Everyday, your top performers are buried under emails, scheduling, and reporting, work that doesn't move the needle. The result? Frustration, burnout, and worst of all, lost revenue. That's why we created the executive's guide to scaling high performers, a simple, battle-tested framework to eliminate busy work and unlock your team's full potential. Inside, you'll get the exact steps top executives use to retain A-players, boost productivity, and free up their team for high-impact work. Stop losing revenue. Go to workergenix.com/growthhub or check the episode notes to download the executive's guide to scaling high performers now. I love it. It's so important now too because, you know, in the past, VAs often were seen as just handing off low-level tasks, but it's changed now because technology does the low-level tasks. And so the VA needs to be the master of that technology to execute all those things.
Mike Todasco 17:37
Absolutely. I use VAs. I mean, Harley, this is how I got connected with you. It was using a VA that I have because I love doing podcasts. I love having these kind of conversations. And so I actually have two different VAs that work at different places to do this and they have different styles and so forth. And like I think if anything, this is going to expand the market for virtual assistants out there. You know, I've been using VAs for the last 15 plus years now. Like even when I was at a big company, I was still using a VA when I wasn't supposed to be, you know, and all these kind of things like there is so much value in that.
And I think VAs are just going to be able to help that much more in the near term by using these tools and just showing people like how helpful they really can be. Well,
Harley Green 18:21
I think you kind of mentioned on about some of the topics on this next question, but, you know, people often see AI as being a replacement for human roles, but many believe and you probably have this opinion that it works best alongside people. How do you see AI enhancing the capabilities of executive assistants?
Mike Todasco 18:37
Yeah, so look, I would say for the foreseeable future, that is the role of AI is they are not going to be replacements yet. In part because when stuff goes wrong and hallucinations or confabulations within AI happen all the time, and these are things that as long as you have a human at the wheel, they're easy to catch, they're easy to identify. Those are going to continue. They're not going to stop anytime soon. And until they do, you want to have a VA there to catch those things, to be responsible, to talk to about that. So like me as somebody who's hired a VA never even sees those errors and things like that.
The last thing that I would want to do if I was trying to hire a VA is just have problems that I have to then sift through and the documents related to travel, related to whatever that a human being would have caught. And it's not worth it just working with an AI agent to do that. Now, four or five years from now, maybe that's a difference. It's hard to say. But at least as of today, it is a tool that is going to be supplementing and augmenting the things that you do and helping what you do better, whether you're a VA or an engineer or a lawyer or anything else that we put.
Harley Green 19:52
And kind of going back a little bit, we mentioned some strategies at the beginning of the podcast episode about how to encourage adoption, more on the leadership side. Leadership plays a huge role in adopting these AI solutions. What mindset shifts do business owners need to make to fully embrace AI-driven decision making?
Mike Todasco 20:10
Well, look, if you are saying that you're an AI-first company and you're not using the tools yourself, then you're just a hypocrite. Like you need to be the person who is up front and just saying, hey, I'm going to actually start with the all-hands with showing you how I've been using AI. And maybe it's elementary. Hopefully it is. Like, you know, if you're leading an organization of 100 people and you are the one that is most advanced on AI and you're the CEO, that's probably a gap. There's going to be just law of large numbers. There's going to be people that know more than you.
And that's okay. But you want to see the leader of an organization doing it first because it will help you. It will help you improve in everything you're doing. Toby from Shopify actually just recently put out an amazing little memo that got leaked was saying how Shopify is going to be using become an AI-first company. Basically, if you ever want to add headcount, you need to prove why this cannot be done with an AI. Like just simple things like that within an organization. And Toby put out this like four-page memo that was sort of outlining that. And, you know, if I'm a leader in a company, I'm thinking, well, what's the version I can have?
And it's available for anybody to see on Twitter if they want to. It's pretty easy to find. It's gotten a lot of press. You know, what's the version of that within your organization? So do it yourself. Don't be a hypocrite and just clear communication on what you want. And I guess the third thing as we already talked about, you need to have tools that are approved, that are encouraged, that are on everybody's desktop, that anybody in the company can use. So if you do those three things, you will become an AI-first organization.
Harley Green 21:45
Now, for me personally, I'm really big on like KPIs, tracking performance of my team and with the business. You know, talking about AI adoption and AI-first company, if a business is wanting to adopt AI more and be truly AI-first, what are some like positive kind of team, you0 leadership type KPIs or other policies or frameworks that you would recommend they start implementing to really encourage the adoption beyond what we've talked about of just, you know, showing examples in like those meetups or from the leadership to really make sure everything that's being done is truly trying to put AI first.
Mike Todasco 22:21
The obvious one's just like usage of these tools. And I am sure all of these tools in an enterprise environment have dashboards where you can actually see who is using it the most. Honestly, I would like start to reach out to those people and those would be the people I'd be on stage and just like how do we encourage? How do we do that? Like I'm actually not as big into KPIs as the stories that we tell and that the stories will really inspire people within an organization more than saying, hey, you have to use it five times per day or whatever it might be.
But you can use the KPIs to help guide that. And so see why the top users of these tools are doing it. Maybe it's for some personal stuff. And you know what? That's okay. Because even if they're using it for personal reasons, it's still going to bleed into their workspace. They're still going to be able to find a way to make it useful, to make it helpful. And I would really showcase that. I was listening to a podcast recently, this week in startups, and Jason Calacanis was talking about one of the things he looks at within his, you know, relatively small organization that he has is since they've been sort of AI-first, nobody's asking for actual more headcount.
And so like really a KPI there, but like when you start to, you know, usually and we've all managed people in the past, we kind of know like, oh, everybody's like, oh, I'm strapped. I have two like whatever it might be. But I think that's a little bit of a tell once you're in an AI-first company when employees who may have been just saying like, oh, we can't get this stuff done. There's too much like, you know, just saying like we need people, we need when that starts to get quiet, that is probably a little bit of a tell that you're actually becoming the AI-first organization because the AI is doing the busy work.
You know, you're using superhuman to help everybody with email and saving the entire organization 30 minutes per person per day or whatever it might be. There's so many tools out there to help increase efficiency and productivity and just the quality of the
Harley Green 24:20
Output. Now, for entrepreneurs and leaders that are a little bit hesitant about adopting AI, what's the biggest misconception they need to overcome in order to stay competitive and start adopting AI? Well, the first
Mike Todasco 24:31
Thing I would say, it's not misconception, but like if you're hesitant, your competitors are not. That's the first thing that I would say. And I think maybe a misconception might be that it doesn't apply to our business. It just doesn't mean anything to us. We're Del Monte and our business is about making banana or whatever, not making bananas, I guess growing bananas and distributing those and so forth. But like I guarantee if I got into the weeds of Del Monte, like there's a lot of efficiency and a lot of things AI can improve in the processes being done across that organization. Here's the reality. Two years ago when I would talk to students, I would say the best job for you as you're graduating college is to learn AI and to go into a company and you will go in there and you will have superpowers because nobody else at the company would know what to do.
That world's changed. I actually don't give that advice to students anymore. And actually what I tell them now is the only job that I know that's going to exist in five years is that of a solopreneur, is somebody who is just a one-person shop who is passionate about a product, passionate about a customer space, builds something for that. And you might have 100, a thousand, a million agents working for you at your direction or whatever that might be. And this is looking a little bit in the future, but I'm trying to kind of say like, okay, if you're 22 now and you're going to be 25, 26, what's the world going to actually look like?
And what I would say to any organization out there like, okay, so say you're a software company, you're you're building something, the AI can replicate your what you've built on a on a moment's notice. Like what is your real moat? If you could have a dozen startups with all these agents competing with you out of college building a similar software or maybe capturing every individual niche that you're targeting, what is your moat in that world? And this is the biggest shift that we're seeing. Like right now where we're at in early April, I am sure every business owner across the world is so concerned about tariffs.
Tariffs. That is all that's going on. That doesn't even matter compared to the seismic shift that we're going to see from AI in the upcoming years. Like sure, okay, spend some time on that. Like yes, I could see how that for many people, especially importing, exporting physical goods, that can be an issue. But the bigger thing out there is AI and what the opportunity is going to be and also what the risks are going to be. So for anybody who's hesitant, that's what I would say. Like, you know, you got to look at the strategy. You got to look at like what the longer term is and kind of just move past that and just say like even if I don't like it, even if I don't understand it, this is just what the world is going to be like.
For the sake of the future of this organization, I need to embrace it and I need to kind of do the steps we talked about before to make sure we're
Harley Green 27:21
AI-first. Absolutely. Well, Mike, you have shared some incredible insights and wisdom with our audience today about AI. If people want to learn more or stay up to date with AI, what's the best way for them to connect with you?
Mike Todasco 27:31
I'm on LinkedIn if anybody wants to follow me there. I do have a newsletter I put out. It's totally free. I don't make money off of any of this stuff. Like that's not my goal at all. I just want to blab about interesting things in the world of AI in a very plain spoken way. Again, I am not an engineer, so I try to bring it down to real terms and make it real practical for folks. So if they want to follow that newsletter, I encourage them to do
Harley Green 27:54
So. And we'll be sure to have that link in the show notes wherever you're viewing or listening to this. We'll be sure to check it out. And if you got value from this episode, do one quick thing. Hit like and subscribe so you don't miss future strategies to help you scale smarter. And do you know a business owner or colleague who could use this? Share this episode with them. It could be exactly what they need right now. And if you're listening on a podcast platform, leave us a quick rating. It helps us reach more leaders just like you. And thanks for tuning in. We'll see you in the next one.