What has stopped the adoption of AI? Too many moving parts. Hear from Ken Mills of Intellisite and Epic IO about how it’s coming all together.
Ken Mills:
Thanks for having me today. My name is Ken Mills. I’m the CEO of EPIC IO Technologies, the parent company of IntelliSite.IO, a partner of Chooch. We’re very excited today to talk about where the industry is going, and where AI is trending and one of my favorite topics are how customers are looking at AI and how they’re expecting us to sell to them and deliver AI in the future. The best way for me to describe this is customers aren’t looking to buy another box of LEGOs. They really just want a hot and ready pizza. So let’s jump in and talk about the AI best known use cases that we all know and love: operational efficiency and really improving manufacturing, industry 4.0, security, whether it be site security or surveillance or anomaly detection, safety, pedestrian safety, traffic flow, vision zero. There’s all these great examples that we love to talk about and provide use cases for and outcomes for, for our customers.
Ken Mills:
We often ask ourselves what is keeping our customers from really latching on these AI solutions and deploying them everywhere possible. Although AI is exploding, as we can see here on the numbers, by 2030, AI for computer vision specifically, it’s going to be a 41-plus billion dollar industry, where AI overall, if you look at all things AI, is going to be over 190 billion in 2025. So, just around the corner, we’re looking at a almost 200 billion market opportunity in across the globe. And really it’s pretty amazing to see what profound impacts AI is going to have. That being said, there’s still challenges that we see every day with our customers and how they’re deploying, or what’s stopping them from deploying AI solutions, whether it be computer vision, data AI, audio AI, or any combination of the above.
Ken Mills:
And, in my studies, in my own experience over the last 20 years, 10 years really, selling AI solutions, these are the three items that typically come up. And there’s a number of surveys out there. And this one had some really good data points around 56% of the people who are struggling to deploy AI, they struggle because they just don’t have the skills on staff to really consume that data and do something with the output of those great models that we love to sell and talk about. So not having the trained resources, not having to train staff to really do something with that data in a meaningful way is a big barrier for a lot of our customers to really adopt AI and really go forward with it on a day-to-day application.
Ken Mills:
Another major example or reason is really just fear of the unknown. They don’t even know what they don’t know, right? They don’t know what they’re going to find. They don’t know if the outcomes going to meet their needs. They don’t know if the ROI is going to be there. They don’t know if they’re going to find something they really don’t want to find. There’s just all of this kind of unknown and anticipation of what may or may not be as come out as a result as of deploying AI. So really just that fear in itself of what they don’t know is a roadblock for customers.
Ken Mills:
But one of the ones that is something near and dear to my heart, and what I’m really talking about today, is customers struggle with finding that starting point, like what AI model do they start with and what outcome do they focus on and what technology do they buy first and what piece of the LEGO do they invest in first so that they can continue to build on top of that and get to where they ultimately want to go. And that unknown, that fear of unknown, that inability to figure out where to start and that just discontent about where they’re going to end up, all those things together, ultimately stop customers on a day-to-day basis from choosing not to do an AI project.
Ken Mills:
And, obviously, in this session and the customers we’re talking to, we’re hoping that they don’t fall in this category and that they’re very eager and ready to deploy AI solutions with us. But we have to overcome these challenges across our customer base to make them feel confident that we can help them solve real challenges.
Ken Mills:
So, when you look at the current state of AI, I hear this a lot in presentations I’ve been guilty of using this in presentations myself, is we are delivering building blocks of tools and functionality that customers can layer on top of each other to ultimately get the outcome that they want. And I think it’s a great analogy, but what the reality of what we’re saying is that they need to figure out where the cloud fits in. They need to figure out where the edge fits in. They need to figure out what dashboards they need, what rule engine, which model, do they want to connect IoT, how do they actually connect to the data, 4G, 5G, wireless, SD-WAN, you name it, right?
Ken Mills:
There’s all these things that have to be it out. And they’re all little building blocks that are added on top of each other one after the other. And, ultimately, you get the full solution. But what I see and find is that we often lay that responsibility on the feet of the customer to figure out how they put this together and which pieces they buy and how they actually fit together and do they fit together the right way and is the end result something that is actually tangible and delivers real value. And, when we leave it up to the customer, we don’t give them all of their directions, we don’t give them all the tools, they often struggle to really put these building blocks together and accomplish what they need.
Ken Mills:
So we also see that customers are really frustrated with a point-solution approach, where we take a AI model-only approach, like “We have the best model. It does all these awesome things, and it’s going to always work 100% of the time everywhere we deploy it.” And it’s not always a single model solves all the problems, right? Usually it’s a multiple model, multiple need situation for most customers. A single model, one-size-fits-all approach is often not the right answer and doesn’t help our customers get where they want to go and help really solve their problems.
Ken Mills:
We also see a lot of siloed solutions where you take that one model and you take it one step further, like maybe a siloed license plate recognition system, where all it does is that one thing. And it doesn’t really well, but it doesn’t play nice with others. It doesn’t integrate with the rest of the ecosystem. It doesn’t allow data to be shared to other applications, to be integrated with other third parties, to really extend outside of the LPR solution itself. What that ends up forcing the customer to do is buy multiple siloed solutions that sit next to each other that aren’t efficiently using compute infrastructure, aren’t efficiently using application resources, require separate licensing, separate trainings, separate support, separate management, and it can get very cumbersome for customers as they layer on all of these siloed solutions on top of each other to get to a business outcome that maybe they don’t even know what they’re getting from when they ended from where they started. It can be very, very confusing for a lot of customers.
Ken Mills:
So customers are really looking for a more integrated solution, a more complete solution that does not require a siloed individual approach. They want a solution that can accommodate future ideas, that’s opened, that allows third party applications and data to be integrated organically with the solution, and really does play nice and well with others in the ecosystem because customers have been bitten by a siloed approach from a company that goes out of business, and then what do they do from there? They’re left with a very limited set of options, and they’ve already spent the money.
Ken Mills:
We’re also seeing a major trend in moving away from silos but also moving into bringing multiple technologies together to solve solutions. We’re starting to really see AI and IoT come together as cohesive solutions where both are adding tremendous value in cooperation with each other, as opposed to being delivered independently. So AI is delivering great models and great insight and results from those models. And then IoT is over here separately delivering data insights and sensor data and real-time data. But, when they’re completely independent, you’re missing the real value that they can bring when they’re brought together. So we’re seeing a significant emergence of AIoT where AI and IoT co-exist and the same platform, the same customer data set, and are really bringing a better together, more rich data story so that customers can make really meaningful decisions with all of the information, not just one subset of the information.
Ken Mills:
Think about your senses. Using all five senses to generate the best result, the best outcome, the best information. By bringing IoT and AI together, you’re using many more of your senses at the same time, generated a much richer set of data and results for the customer to make a decision, which is ultimately why people invest in our solutions in the first place.
Ken Mills:
Ending this all together is why do customers want pizzas and not LEGOs? If you think about it, when you are a kid or you have kids, or both, you have this grand idea when you see a big LEGO set of whatever LEGO character of your choice. There’s, I think, every option you can think of, LEGO’s built something around it. And you can go buy that 100-piece, 3000-piece, 80,000-piece LEGO, whatever the case may be. You can spend a ton of time putting it together, trying to figure it out. You get frustrated. You may quit. It may never get put together. You lose LEGOs. People step on them. That’s a whole nother problem and pandemic we need to solve. You know, stepping on LEGOs randomly in the middle of the night, but really struggling to figure out how to get the end result of that thing you saw in the store, having to be responsible for putting it all together.
Ken Mills:
And then, when you finally get that LEGO together, you spent so much time, so much energy. You’re afraid to use it. You’re afraid to play with it. And it’s just this fragile thing you sit on a shelf’s somewhere and you hope no one touches, no one breaks, and you never have to rebuild it again. That’s not what customers want. They don’t want that from their AI solutions. They don’t want it to be so fragile, so complicated, so time consuming that they don’t get to actually enjoy the outcome that they anticipated when they made the decision to purchase or deploy AI.
Ken Mills:
They really want a pizza, right? When you are hungry and you’re thinking about, “What am I going to eat tonight?” You don’t really want to cook. You get on the internet. You get on your phone. You order pizza. It comes to your house. You open the box. It’s hot, ready to eat. You pick it up. You eat it. You enjoy it and you’re done. No fuss, no muss, no effort. You got the whole solution with all the ingredients nicely put together, package delivered and with an exceptional outcome every time. That’s what customers looking for. They don’t want us to sell them LEGOs. They want to buy pizzas.
Ken Mills:
I leave that with you. Stop selling LEGOs. Sell more pizzas. I really appreciate your time. I hope you have a great rest of the event. Thank you so much for allowing us to be a part of it.
