AI Adoption Myths Debunked by NineTwoThree AI Studio CEO Andrew Amann

Andrew Amen is the CEO and co founder of 923 Studio. Join us as Andrew takes us through the fascinating journey of 923 Studio, a development and consulting agency renowned for solving AI problems for major clients like SimpliSafe and Experian.

From its roots in a mobile app and web development in 2012 to pioneering AI solutions since 2016, Andrew shares the evolution and impact of AI, the challenges faced by big companies in adopting this technology, and the critical role of customization in AI implementation. Plus, get a glimpse into Andrew's entrepreneurial spirit as we delve into his experience launching 14 startups and the importance of having the right team for business success.

As always, if you found value from this content, please like and subscribe. Hello everyone, and welcome to another episode of the Beyond Fulfillment podcast. I'm your host, Dave Gulas, and this week my guest is the CEO and co founder of 923 Studio, Andrew Amen. Welcome, Andrew. Thanks for having me, Dave. It's a pleasure to be here. Yeah, appreciate you taking the time, if you could, for everyone. Can you tell us what 923Studio is and how it got started?

Sure. So we're a development and consulting agency. We build products for SimpliSafe, Consumer Reports, Experian, some of the largest clients. We basically go in and solve their AI problems. So they come to us and say, you know, we have an initiative to install AI into our systems and we don't know how. So we do workshops, we do consulting, and then the biggest part of our business is going in and doing the development and the product release from 0 to 1.

Especially for big companies, they want a new initiative or new product so we can get in there, get our hands dirty, figure out how to build the correct product to create efficiencies. And I like to say what we're doing now is increasing revenue per employee because we are truly creating revolutionary returns on investment just from increasing that revenue. We got started in 2012. We were a mobile app and web agency for many years, and then our first AI or venture into AI was in 2016.

We worked on a baby monitor app that had some infrared sensing and some machine learning involved in it. And I would say we did a lot of machine learning from 2016 to 2022, and then a lot of generative AI from 2022 till today. Okay, all right, so a lot to unpack there with, with the journey of this company. So AI itself really didn't become mainstream till, you know, let's say, the beginning of 23 or so and most people didn't know what it was prior.

But you mentioned you were working on it since 2016. Yeah, the act. Yeah, it's actually the second wave of AI. AI 1.0 is got everyone addicted to social media. So the machine learning algorithms and the feeds that we all use inside of Facebook, Netflix, any of those places that are doing recommendation engines, that is really the first addicting AI model that has been deployed at massive scale that, you know, 99% of humans use that have connection to the Internet.

So we're really in version two of AI or wave two of AI. Okay. And at that time, 2016, when you're working on this initial product and the baby monitor and all that, I mean, did you and your team, did you see. Right, because you look at it now and how, how, how big it's got and all the different uses, applications and whatnot. Did you, did you see it getting this big, you know, this soon? I guess, yeah. There as, as it's a. It's like a family, Right. So AI is the top of the pyramid.

It's like calling fruit. It's calling Apple a fruit. Right. So calling machine learning, natural language processing, data science, all of that is part of the AI family. In 2016, we were doing majority machine learning, so time series, a lot of data, trying to figure out exactly prediction models. So we predicted houses, we predicted when babies were going to wake up, we predicted a lot of things inside of manufacturing plants.

All of those are machine learning based systems which I still call AI because it's part of the family of AI. What changed in 2022 is the invention of, or not the invention, but the adoption of transformers and those transformers, then creating these large language models, and then us adapting to the large language models with our current team. So NLP engineers we've had since 2016, very, very good at deciphering how large language models work.

And so they had six, seven, eight years of experience with that. So by 2022, we already had projects that had been released and on the market. So when the hype came, we were very well positioned to use those engineers and use that expertise to build a lot of the ChatGPT stuff that was coming to the forefront. Winning deals is as simple as the sales tools that you use.

Pipedrive's easy and effective CRM lets you track your sales pipeline, manage deals and automate the whole process so you can focus on one selling. Join over 100,000 users worldwide, driving business growth with Pipedrive. Customize your pipeline to match your unique sales process, sharpen your Strategy with real time performance insights and centralize Data from over 500 sales tools that you know and love.

Make smarter, faster decisions with pipedrive's AI sales assistant and win up to three times more deals with smart automations. Getting started takes minutes. Try pipedrive today for free and unlock real results from day one. And like you mentioned, you're well positioned because you'd been in the space for a number of years.

So when that initial boom hit and everyone really wanted to get into AI, what, what was it like in terms of, was it not so crowded in, in your, in your area of the market to where you could just. Because, you know, just, just win those large accounts that you mentioned earlier? Yeah, it's actually surprising. I would even say up until today it's still not a, you know, it's not like a hype for us. We still have a team available in January. We're not overbooked.

So customers and clients, when they hear that they need to install AI or they hear of another company using AI, they think, great, what product did you use? Can I use that myself? And the reality is they can't. It's something that is very custom based for a company. But also the efficiencies that you gain from AI are because your team has adapted to using the correct systems and the right calls and the right summarizations and using those systems that were built specifically for their company.

And so they're experiencing the rewards and then shouting from the rooftops how great the return on investment are. So to hire an agency, you're getting a kind of like a head start, but also a quick iteration around how to do it properly. And it's very hard to convince CEO of a 300 million company to jump into AI when they have engineers in house that are reading ChatGPT and think they can do it themselves.

It's also very hard to convince a cto of a $500 million company to say, I don't have anyone on my staff that can do this. Why don't I hire outside? So all of those shifts are happening and they're happening slowly inside of companies.

While AI on the surface and what we see on Twitter and the news is happening very rapidly, at one point this will converge. And I think once the wave happens of everyone saying, oh great, I understand that I need an agency to install this, who can I go call? Our phones will be ringing off the hook. But still to this day, it's still a crawl for a lot of companies. And CEOs really trying to decide what the Right move for them is specifically in their company. Okay, interesting.

So you're saying it sounds like larger companies too are slower to adopt the agency model just because of it's not the status quo and you know, they're not using their internal resources that they think should be able to do, do this type of work. That's correct. And it's a different way to think. I assimilate it to the cloud transition. And if we remember that in 2012, everyone was trying to get their information to the cloud. Whether you needed to or not was a different decision.

But the entire move was you should be having information in the cloud because it's accessible by mobile phones and tablets and the rest of the world. And for many organizations, that was a great move and great transition. And that one year promise turned into 10 years of pain trying to get all of their local information off of their machines into, into the cloud. I think those CEOs still have a little bit of those scars.

And you know, bitcoin came and went and these, these trends come and they're looking at AI being like, oh, here comes another thing that I have to then immediately jump on. But this one feels different. The return on investments we're, we're providing to companies are happening in months instead of years. And so we are able to tell them these stories, but they're like, yeah, but that was that company, not mine.

And it's still when you have to educate, train and then make the purchase, it's harder than them just coming to you and saying, take my money. And so that's the transition we're dealing with right now. And we are fine educating and training. That's part of our business model. But it's just not as easy just to say, let's go today. Okay? And so with AI and it's been mass adoption, let's say two years or so.

I mean, what do you see some of the biggest mistakes that companies are making in terms of when, when they're implementing AI into their operations. We've had a handful of companies that we approached in 2022 that said, and we said to them, this is how we, we gave them a proposal, we gave them how the roadmap would look like and how we would solve their, their current problems or their efficiency problems. And they said, we'll take it in house.

Fast forward nine months, we're getting those same phone calls from those companies saying it's not working. What are we doing wrong? And so we're able to see ahead of the curve and we're Also able to use these systems not based on the foundational releases. So, you know, ChatGPT 4.0 or 5.0 or 6.0 comes out, it doesn't matter to us. We're building these, these internal efficiency engines that are really solving their problems at the core and then using the large language model as a tool.

And so if ChatGPT 5 comes out, we just swap it with the API calls. There's no re engineering on our end, there's no retraining. So I think a lot of the mistakes that people are making is one, not starting and not getting that knowledge base inside of their company so that they can have a communication with the actual data that they've housed for many, many years.

And the second one is not having teammates or product manager or somebody on the team that is just going to advocate for this and be the champion. When we deploy these systems, we can tell you the day we release it if it's going to be successful because the company has hired or has somebody in the company that is an advocate for making this system work. McKinsey put out a great report that said the adoption to AI is 80% of humans and 20% of the technology.

And that means that the processes, the procedures, the way you think about doing business needs to change. And if you can do that, you'll see massive returns. I mean, we have massive returns in the tens of millions of dollars from $200,000 investment. So you can see those, those returns. But you have to truly have somebody on the team that's going to stand up and say, this is the new way of doing business and oh, look at the effects that it's having.

Okay. And brought up another question too, on the topic of AI being that it's so new and like you said, it's such a dramatic shift in thinking how, from what you've seen in terms of the education system, how schools are looking at AI, whether at the, you know, you know, the graded level or up through university. How do you see that and you know, as far as how, how they're. Because I've heard mixed things from different people, but I mean, you're much more familiar with that area.

How do you see how they're approaching that and what could be better?

Yeah, I'm, I'm, I've been studying this now for the last few years. We have a few clients in the education space and I also have kids that are 8 and 6. One of the biggest changes in my kid's life is I don't think there is A technology that has been as revolutionary for the education system as Condomingo is. The problem is people listening to this podcast and even yourself, might be the first time you've heard of it.

It is the adaptation of ChatGPT for a kid to have a conversation with something that is 100% patient all the time. It's also very knowledgeable and it can keep on task and on, on goal oriented, you know, vision, no matter where the conversation lies. And it's 100% safe. I won't say 100% safe. I've tried to break it a few times and it does say, like, hippopotamuses are similar to giraffes because their neck size is the same. So I've gotten it to forcefully say something that is just outrageous.

But you also want to make sure these systems stay, stay, stay safe for kids. But Kanamingo itself is the first step towards an AI tutor for every kid. And it is something that can start remembering conversations, it can start remembering grade levels. And as it asks for specific feedback, which it's doing better and better, it will say, I was asking it the other day with my daughter, I was like, how many planets are there? And it said, eight planets.

And then it asked her to name them, and she would name them one by one and she would try to think and remember, but at the end of it, she missed one. And so she said, what planet did I miss now? You know, if there was me, that was 10 minutes of a conversation of her thinking and asking. Adults, you know, we get impatient, we get a little nervous, and we want to make sure our kid knows the information. So we just kind of say the answer. But with something like Condomingo, it can be trained.

It can be through a process in which is still allowing the kid to go through their learning style at their pace. And that is the first step I've seen to an AI for every student. Fast forward that two to three to five years. I think it's ridiculous that we have an education program where every student sits in the class based on their age, regardless of their education level. I mean, my daughter specifically, she's in third grade.

She finished her third grade math program that's on her computer and they said, start it over instead of going to fourth grade. I think that progress is detrimental to kids because the ones that do want to learn fast and can learn fast should be able to go to the next grade level and continue on.

While that teacher in the room might not even know or teach that grade level, they should be able to babysit and understand that this kid is progressing because a computer is taking care of their knowledge, not the teacher. So I think that the advantage like that, the thing that AI brings to us is that AI tutor for every student.

Are we going to allow this AI for every student to adapt and replace how the teachers put their curriculums forward in front of a student, hoping that all students in that grade have the same learning paradigm? But that's not true. We all know that if anyone has kids, every kid learns differently.

So I think it's up to us as the adults to start looking at these systems and saying it is possible for a kid to go grade to grade to grade to grade, as long as the computer is keeping track of that progress. Interesting. Okay, and so with everything you're doing in the field of AI and whatnot, where do you see this going in terms terms of the next five years, in terms of some of the things like you mentioned, what potentially could be possible within education?

It's very interesting point, but just in the business sector, what type of things are we going to be seeing with AI in the next five years or so? Yeah, in the next 12 months, if Apple can get their act together, we're going to have an AI assistant on every phone in our pockets. That AI assistant is going to start to be used more and more because it's going to be able to retrieve answers better than Google ever could. Whether it's right or wrong, it's a different conversation.

But humans are incredibly lazy. So if you're able to just shout into a machine anything that you're thinking, acting, doing, wanting to text, any actions you want to take, and it responds generally well and it can start doing tasks for you like setting calendar appointments, setting doctor appointments, calling doctors, going out and calling the insurance company and making sure your insurance plan is renewed.

Once it starts doing those things and we're very close to having it complete those tasks, I think what we're going to see is people getting addicted to using their AI voice assistant. That leads us down a path in which Apple and Google have an incredible advantage because they have hardware tied with software which has a large language model on them, whether it's in the cloud or local doesn't matter yet.

But that information is now mean that every search you do goes through Apple or Google and it becomes dangerous or very rewarding. Because right now I have a browser. I use Microsoft Edge. I then use Microsoft Edge to use a search bar which searches the information inside of Google. After I retrieve that information into Google, I go to Another site, let's say Airbnb, I use that search bar to then find a house.

So from beginning of my search query all the way until finding a house to rent, Google did not share the information with Airbnb of where I came from, and Airbnb didn't share the information with my phone specifically about me, about which house I'm buying. We're going to have to start giving in to the fact that the AI search assistant that's living on our phone, whether it's Siri or Gemini, is going to know everything we do with all of our apps.

And therefore there's going to be one place of all of my knowledge. How are we going to protect that? How are we going to control that?

And maybe Perplexity and Claude comes out with their own system or their own version of this, but I think the future does lie in that search bar that lives on the phone and starts utilizing all of our tendencies and starts predicting for us what we will be searching and who controls that, who owns that is going to be the fight for the next five, six years inside of the big tech giants. Okay, another point you mentioned there, humans are incredibly lazy. And I've noticed that myself.

Just dealing with numerous people out there that, you know, with AI, like almost AI is making people lazier with what it can do, where people are just solely relying on what the AI can do and make maybe taking, oh great, I don't have to think anymore, I'll just paste it in AI and it's done. I mean, how big of an issue do you see that within the workforce, so to speak? And what are some of the consequences of, I guess, that kind of behavior?

I think there are two companies that are advocating for that, which is Apple and Facebook. And I think there are two companies that are advocating for the intelligence of using AI, which is Google and Microsoft.

And what I mean by that, if you look at what Apple and Facebook, what they advertise to and who they advertise to, they try to get you addicted to a piece of hardware or piece of software that is more dedicated to lower income people that maybe can't afford a high end phone or a high end goggles that you put on your head, but that addiction of if you wear an Oculus Rift and you go travel and see the Egyptian pyramids through the oculus rift for 50 bucks, that is something that you no longer have

to do in real life. And that addiction of being able to visit places and see places and interact with pieces of hardware while sitting in your own Home is going to be very difficult for lower income people to get away from because the cost difference of just doing it in your house compared to doing it in real life is going to be catastrophically different.

The other challenge that I think Apple plays into, if you look at where they installed Apple intelligence, first it was the text messaging, which is a shame because the emojis and the emoticons and whatever Apple calls them nowadays is revolutionizing how people communicate with each other. But it's also really just taking a stab at the fact that we don't have anything innovate like Apple doesn't have anything else innovative to go after besides the text messaging and the emojis.

Well, why is that? It's because that is how people are communicating with each other and that is where people are spending most of their time. They're addicted to those blue bubbles, those green bubbles. And if they can get notifications on their phone from friends or family members and they can make sure that they pick up that Apple phone more and more throughout the day, they can keep them addicted to those devices. Google and Microsoft are playing a different game.

They're trying to take the devices out of your pocket and trying to get it into voice and trying to get devices not to be addicting to interact with. And so you can see Microsoft and ChatGPT. It's having a conversation, but it's also creating that conversation in a way that I don't have to then have that Bing or that notification come back to make it work.

And Google specifically, they're talking about making devices at Google Home and using Gemini, more of a voice trained model, and being able to press that button on the side of your phone to activate that voice model so that you can have that conversation and not really focus on the addiction model, but also use that AI as an assistant, use it as a guide, use it as a Sherpa to find information, to help you with information.

And so I think that's the veering of directions we're going because the lazy humans are going to fall into the trap of Facebook feeds and traveling using Oculus Riffs and, you know, scrolling Facebook. I think the average now is almost 50 minutes a day per American. And watching Netflix for way too long. Like all of those addiction traps are AI based. It's the algorithms running in the background to make sure we stay on those screens longer and longer and longer.

Google and Microsoft don't have those products and they don't have those products that keep you addicted to screens for a long period of time. They're more focused on the business world. Right. Gmail and Calendar. And Microsoft's focused on Microsoft Office 365. So they have different goals, different metrics. But I do, if I had to pick, live in the world where I don't have a device on me and I just have an assistant that I can talk to. Okay, interesting.

All right, so we spent a lot of time talking about AI and whatnot. Let's just go over your journey a little bit. So you've launched 14 startups? That's correct. So how did you initially become an entrepreneur?

I don't think I realized I was an entrepreneur until after college. I started working. I was a nuclear submarine engineer in working as a civilian for the Navy. And I quickly started developing products for them that allowed me to be in rooms that younger person typically doesn't be a part of. And I was designing these machines or these ideas of how submarines can be ready for the next generation of warfare.

At that same time, I was starting to dive into software and starting to figure out how software kind of integrated with the hardware itself. And I invented three chips that you put onto product supply chains inside of manufacturing plants, and you have a wi fi bridge that communicates with those chips. Well, the companies that I worked for owned the patent to those three ideas.

And I traveled the world trying to install this patented supply chain process into all these different manufacturing plants in France and England and Scotland. And I was traveling to do this, and I realized it would be nice to own my own piece of technology and own the entire product that I'm thinking of and trying to install. And so I switched over at that point to starting my own company.

But I truly think the idea of entrepreneurship really wasn't something in me until after college when I realized that I can do this, you know, and I had enough ability to take that risk and enough ability to know and belief in myself that I can start a business. But it wasn't something that I don't think I was born with. I think it's something that I always innately knew how to do.

And then by the time I started creating those chips and running around the world, I. I was willing to take the risk and be my own boss. So being so young and designing products like that for the Navy and then, you know, traveling and, you know, installing the chips, I mean, were you like an exceptional student? Did you know you had this type of inventive, I guess, skill set? You know, growing up and through school. I played with legos and connects a lot.

So I definitely had the engineer in me, even through school and stuff, I definitely knew how to build and put things together. I think where my superpower comes in is really when I got into supply chain, I truly understood inputs and outputs, and that maybe came from education, but that also came from mathematics of just knowing what things really relate with each other and which things don't. I think that's why I had the opportunity to build these design parts.

It's because I looked at the world a little bit differently and I allowed the team to explore ideas that they maybe didn't explore before. So I think if you mix design or critical thinking design with supply chain concepts, every business, I truly think this.

I know I'm manufacturing biased, but I think if you learn supply chain theories and you learn about the theory of constraints and you learn about bottlenecks, you can be a consultant for a lot of businesses that haven't figured that out yet. Whether they're in manufacturing or not, it doesn't matter. It's. It's an entire process of how to successfully move a business without hitting roadblocks and move it forward profitably, because manufacturing is always running on thin margins.

So I think if you take that skill. And it was on the job training that I got from, you know, the age of 2018, I was 18, 22 years old all the way until I was about 27, was on the job training, being a supply chain analyst, working on the floor and just hanging out with manufacturing for those years every day, mixed with this desire to build products that people cared about and design.

And design thinking in manufacturing, I think, is a new concept that they still haven't really uncovered or really dove into. And I think I was the right mix of design theory and manufacturing theory. And putting that together from on the job training, I think, is what allowed me to be this unique entrepreneur. Okay, and can you just run down. I mean, so 14. So you started when you were in your late 20s, sounds like. And then how.

Like, give us a sense of how quick you moved on those and why you did so many different ones. Yeah. So 2012 was the first product that was not owned by somebody else. It was called the Nigo is a digital business card. We built that until 2016 and sold that to a Canadian company. At the same time, people had started asking us to build other apps. And that's when we got into AI and started building for Derel, which was a baby monitor, the first infrared sensing baby monitor for any camera market.

And as we sold the company in 2016, that's also when I finally left my, my nuclear jobs and went full time. So from 2016, 2017, 2018, we started building products for other people and we started doing very well in building those products. So we used the profits to build our startups and it was basically ideas that we had. We had a bench of engineers and designers and qa and they weren't doing anything half the time because, you know, we had a lot of work.

But we also had these, these periods in which we couldn't win new projects. And so we used their time and converted that to products. And so one grew to two, which grew to five, which grew to 10. And shirt surely or later like we had 14 products and I had five CEOs that were. I was managing and supporting and trying to help at the same time as running the agency, which was the profit engine of building all of the startups, like providing the money for that.

And so I was really six CEOs, me being one of them, trying to navigate 14 products. It was chaos. It was a mess, it was not profitable. But it was a lot of fun and it gave us the framework of having a lot of shots on goal very quickly to be able to do what we do.

Now we can walk into any business in a lot of industries that are looking for software consultancy and solve their problems because we've probably seen it before, built a startup in that space, studied it, are very educated in how business models work, and we really can speak from experience. So from 20, what would that be? 2018 to 2022, I would call it a venture studio. We ran that venture studio, a bootstrapped venture studio because we didn't have any, any funding.

And then in 2022, we shook hands, me and my co founder, and said, that's it, no more.

Let's just build for other people, let's build for clients. And so we started doing that in 2022, install the OS into our agency. And we've been on the Inc 5000 now for four straight years. Okay, so the Inc 5000 with four years is with this agency here at 923 Studio. Okay. All right. And love that you brought up EOS. Right. So just, just speak to everyone for a moment about that. And I mean, their model is having the right people in the right places within a business.

How important has that been within your agency? It is probably the number one reason we're successful is. And it's not eos. Like the idea of EOS is what's more successful than EOS itself. It's like saying you Know, running a business needs a profit and loss statement in an accounting book. Well, of course, to be successful you need to do accounting and you need to do profit and loss estimates. But it doesn't matter if you use Xero or QuickBooks or your own personal bookkeeper.

It's the idea of using the system itself. And so EOS is one of the many systems you can do. Ours is run by Gino Wickman. We have an implementer that we pay a lot of money to who's well worth it. We do four quarterly, what we call leadership meetings, and those are planning meetings that set us up for each of the quarters. And then our entire company has rocks. So as as a buy in to eos, you end up having these quarterly rocks that you commit to the company that you'll complete in 90 days.

And so the entire process, it's, it's a cadence machine. It works off of the same seasonality that we have as humans with Mother Nature. And so every quarter you're going through a different set of goals, a different set of metrics, and you have these scorecards that you're measuring every week. But you think of it as setting up the cadence for your company, for everyone to start rowing in the same direction.

When something gets off kilter or something doesn't work, the entire system realizes that we need to fix that part of the business. And so the advantage of EOS is putting visibility across all the departments, making everybody accountable and then putting the right people in the right seats to kind of row in the same direction. And it's been insurmountable for us. It's one of the greatest achievements I think we've done as a company.

We went from not having EOS to graduating EOs in two years, which is very short. And now we operate fully remote, like we're a remote team. We operate fully with EOS quarter to quarter. And our implementer now just does our yearlies with us. So it's been a very, very great achievement for the business and very successful for us. Okay. All right. And Andrew, if people want to get in touch with you, learn more about your agency and what you do, what's the best way that people can reach out to you?

Sure. So I'm very busy on LinkedIn. I write every Sunday night. Put those articles out five times a week and you can contact me through messaging there. I check the messages. But also if you do want to contact me or my business, It's Andrew@923co and it's spelled out with the letters N I N E T W o T H R E so I appreciate you having me, Dave. This has been a lot of fun. Thank you for all the questions. Yeah, thank you, Andrew. And we'll link all that in the show notes for everyone. Sounds good.

That is all the time we have for now. We will see you next.