On this episode of the AI For All Podcast, Frank Pica, Co-Founder and CEO of Native AI, joins Ryan Chacon and Neil Sahota to discuss AI and digital twins. They talk about AI in marketing, creating digital twins of customers, generative AI, bringing AI to your business, data and privacy, social media, the impact of ChatGPT, and the future of AI and digital twins.

About Frank Pica

Frank Pica is the Co-Founder and CEO of Native AI. Pica brings over a decade in the AI space, specializing in intelligent marketing and advertising. Before founding Native AI, Pica launched ADYOULIKE, a contextual targeting AI solution and Decide Technologies, a marketing intelligence company. He also served as a board member at Vertebrae prior to Snap, Inc.’s acquisition.

Interested in connecting with Frank? Reach out on LinkedIn!

About Native AI

Native AI is an always on market intelligence platform that helps businesses understand, innovate, and create ideal experiences for their customers. Native’s proprietary digital twins use generative AI to create clones of a brand’s target customers and consumers to enable real-time engagement and consumer insights.

Key Questions and Topics from This Episode:

(01:16) Introduction to Frank Pica and Native AI
(01:33) AI and digital twins
(03:38) Digital twins of customers
(05:40) Where does generative AI fit in?
(07:12) AI and digital twins adoption - what's changed?
(09:48) Trust in AI
(10:48) Bringing AI to your business
(14:16) Data and privacy
(20:03) Social media and data collection
(23:26) Companies paywalling their data
(26:29) Open source versus proprietary AI models
(27:40) Is crowdsourcing the future?
(29:45) How accurate does AI need to be?
(30:40) How has ChatGPT impacted business interest in AI?
(33:47) Why was ChatGPT so successful?
(34:57) Future of AI and digital twins

Transcript

- [Ryan] Welcome everyone to another episode of the AI For All Podcast. I'm Ryan Chacon with my co-host, Neil Sahota, the founder of AI for Good, AI Advisor to the UN, anything else I'm missing there, Neil?

- [Neil] Ryan, I know you're picking on me, man, because the UN started calling me the Godfather of AI for Good.

- [Ryan] We got to make sure we mentioned that, of course, of course. I also have our producer, Nikolai with us as well.

- [Nikolai] Hello.

- [Ryan] All right. So today's episode, AI and digital twins. Digital twins have really emerged lately as a useful tool for businesses, providing insights, foresight with unprecedented speed. So digital twins are now being used in a lot of topics we've talked about on this podcast before, especially generative AI to make clones of target customers that businesses can actually survey and receive unbiased feedback in minutes, much faster than they have ever before. To discuss this, we have Frank Pica, the CEO and co-founder of Native AI, an always on market intelligence platform that helps businesses understand, innovate, and create ideal experiences for their customers.

Frank, thanks for being on the show.

- [Frank] Thank you guys. Excited to be here.

- [Ryan] Yeah, it's great to have you. So to kick things off, AI and digital twins, what are they in the sense of like, how do they work together? What value are they providing working together? Maybe just start there to help our audience really understand what we're going to dive into a bit further.

- [Frank] Yeah. So this vast world of AI can mean a lot of things to a lot of different people. For us, we are specifically focused on using natural language processing and the generative AI. So combination of those two presents something quite beautiful, especially as it pertains to customer and consumer research.

But as you pointed out in our tagline, it really allows you to take insight and truly turn it into action. It's a tagline I think we've used for years and years but wasn't truly possible to automate. And luckily this beautiful wave of generative AI is really bringing that to life. Digital twins historically used to create a digital twin or a digital replica of a warehouse, supply chain.

It's been extremely useful in scenario modeling to understand where things might fail, where things could be optimized and improved. We talk about it today as creating a digital twin or a replica of a human being. And that beautiful combination of artificial intelligence with cloning an individual digitally truly allows you to bring them to life for, again, everything that we talked about earlier, innovation, to improve marketing, and even test hypothetical scenarios, a change to copy, you can see your marketing messaging and so. That marriage of the two is what Native is best at and what we've brought to market in the last few years.

- [Ryan] Yeah, it's very exciting stuff. We, obviously our other podcast, IoT For All, digital twins is a very common topic in being able to replicate hardware devices, those different environments that deployments are going to be brought into. So when it comes to replicating humans, potential customers, how is that done?

Like what is happening? How, where's the data coming from, how is this different than when we're thinking about it in the sense of how it's being used in IoT with replicating and making a digital version of hardware and things for the same kind of purpose.

- [Frank] Yeah, I think there's a lot of parallels, which is why we chose and coined the phrase, digital twins of customers and digital twins of consumers. The big differentiator, instead of looking at it as a physical asset, you have to look at it as it's a cognitive asset.

And so where IoT can collect tons of information around a warehouse, right around physical objects throughout the world, we look at that very similarly, right? But from a lens of like, how do we better understand humans and their behavior and their attitudes? And a lot of the data that we've collected over the last four years has been just that, it's how are consumers engaging with products and services in market, types of attitudes, right?

And behaviors, so they, you bring to the table during those experiences and to simplify our backend of the platform and in our technology, we look at it in two large buckets, third party data, which again, could be things like product reviews, it could be feedback or commentary left around a product or service or brand that is left on social media.

And then we also utilize first party data, which you could think about a company or brand storing in their CRM after a conversation or perhaps even soliciting feedback around how the consumer's experience was around that product. And we blend those two datasets to create that amalgamation or digital twin or replica of that individual.

- [Ryan] And where does the generative AI, how does that fit in? What's the that functionality look like? Where is, where, what is that kind of allowing, I guess, organizations to, what are they, what is that allowing them to do with this kind of approach and tool?

- [Frank] Let's think of it from like a, maybe, illustration of a brain, right? You have tons and tons of information in your brain stored at any given time. NLP or natural language processing allows you to take that kind of raw, unstructured data or thoughts and generate insight from it.

Generative AI really gives us the mouthpiece, right? So we have all this information, all this data. Now, by applying generative AI, we can actually get it out of the brain, right, and into the mouth and speaking back to our customers. And yes, it can generate insight very flexibly. So, you may have a question for Neil about what he ate over the last year, right?

And Neil can regurgitate that. But also it allows you to ask them questions as like how they feel they should be marketed to our messaged. So we understand Neil's needs or demands or desires. And we can now ask him and his digital twin like what their preference or tone is as it pertains to our marketing messaging, or maybe it's a flavor of product that we have in this research and development cycle. So I like to think of it as the mouthpiece and getting you to that action. That's what generative AI does for us.

- [Neil] I'd like to get your take, Frank, on this, because this combination of like digital twins and AI has been around actually for a while, like even six, seven years ago, we were working to, we were doing like digital twin farms and using AI to figure out different crop rotations. Why has it taken so long?

Because I really haven't seen this kind of catch fire, so to speak, except maybe the last seven, eight months. What's changed? I get the data stuff. What's changed with the business thinking that they're now really into this?

- [Frank] It's a great question. I wish I would have had the answer 10 years ago when I was also speaking about this and people thought I was nuts or off my rocker. I think two things happen from a macro level. One, the technology has gotten much better, right? When I was working alongside IBM with Watson back in 2016, 2017, I thought the technology was fascinating just because it could understand sentiment, right, of text or an article. And I thought that was huge. Little did I know, right, how accurate and how lifelike it could get. And I believe it was, I truly believe this, without ChatGPT, I don't think we would be really even having this conversation today, right? I think it would still be in this idea phase.

ChatGPT, right, put it in the hands of everyday average users, and that's really shaped the way businesses, I think, are approaching, right, strategically, this technology. And I don't need to regurgitate the stats, I'm sure you guys have talked about it all the time, but what we've seen just in terms of lifts in terms of interest from customers is a complete turnaround, right?

Over the six, last six months, so last year, all my time, resources, capital spent educating the market, trying to teach people what was possible with the technology. This year, everybody believes what's possible. Now it's, okay, how do we do this properly? How do we regulate it? And how do we manage things like privacy if we're going to use digital twins or utilize generative AI for our customers or people?

So, I think those are the two biggest tectonic shifts that really happened over the last year that without them, we wouldn't be talking like I said today.

- [Neil] So one, I got to give props to David Lynch then because it sounds like until people could do some like mundane, routine tasks, you couldn't believe it, which I totally agree with, which means David Lynch was right. There's beauty in the mundane. But second, it sounds like there's a little bit more trust in the technology, that they're willing to explore, experiment, whatever the right word here is, with what you're doing.

- [Frank] I wish it was a pure trust.

- [Neil] Open minded?

- [Frank] I think we do see trust at the user level, right? You're going to have your early adopters, people who understand the technology and who do trust the outputs because they've already seen it work for them in a variety of use cases. From a corporation's perspective though, you've got to be careful that taking trust as a table stakes, right?

I feel like that's actually the stage we're at where they're willing to pilot, they're using our third party data, data that's not theirs to essentially get their wheels spinning. But this idea around first party data and protecting their customer's privacy, that's our next phase, and that's what I think about when you start talking about trust.

- [Ryan] I wanted to ask you, Frank, if I'm listening to this, learning about AI, and trying to understand how this, how digital twins and a lot of stuff we've talked about really can play into my business that I'm running, or we're looking to bring this type of tooling into the organization to help us be more efficient, help us better understand our customers, maybe do better research, what are some of the leading use cases companies are adopting this kind of technology for? And I know you've taken us through some examples like using Neil to understand like what, what he's eaten, how he, what his thoughts are and those kinds of things. But if I'm trying to understand how this can be applied, what do I need to know about adopting a tool or a solution like this?

Do I need to already have lots of data? Do I, how am I going to collect that information about these customers? Is it something that you all have access to at a large scale, and we can pick and choose what profiles are available that relate to our customers or how do you approach that or how should people be approaching that or thinking about it?

- [Frank] Yeah. We can talk about this two ways, but I'll let you choose. Broad stroke, there's a bunch of low risk, low hanging fruit opportunities for generative AI. And I'm sure you guys can guess what those are, but let's just talk about something like creating copy for advertisements or for marketing.

Talk about, no offense to copywriters, right, like, it is not an easy task, but it probably shouldn't be as costly as it is today, and AI can create a hundred outputs for every one output a human could. And so as long as that's peer reviewed and brand safe, that would be a low risk low hanging fruit endeavor for any business out there to improve their business and really get their feet wet.

We take a similar approach to our applications towards market research, marketing, or even like product development. And so for us and for businesses, a low risk approach could be, hey, we don't need any data, right? So no requirements upfront. Other than here is the products that we have or services that we offer to the market.

And then this is our category here, our competitors. And just from that alone we can actually create digital twins of their market and their target audience and then go provide insights or go provide actions that they can take to reach that audience better. Or provide a better service or product to that audience, right?

That will result in higher loyalty or more sales on the shelf. And those to me are like the low risk approaches to testing out the technology. And then phase two, if I'm a large corporation, that's when I believe, right, we're getting into how do we blend our very valuable first party data, right, with these third party datasets to get a more comprehensive or what I like to call a more articulate digital twin that can actually provide my business with further action, a more accurate action. And so when you're getting to that point, maybe a little bit higher risk because you're dealing with things like privacy or dealing with, again, essentially regulation forming right before our very eyes.

You're also steering your business in expensive directions if that's incorrect, making new product for a target audience that doesn't exist or that lied to me via an AI hallucination. Those things become a little more risky for a business. And so that's how we've approached the market, and I think it's the correct way, right? Don't jump in maybe with two feet, but at least get a leg in there.

- [Ryan] You mentioned privacy a couple of times now, and I wanted to ask, I wanted to see if you could expand on kind of what is being done to help with customer privacy when you're trying to get that data from your customers to be able to bring into something like this and mix with a third party data to, like you said, get a more articulate answer or more articulate kind of response.

What needs to be done or how is that being taken into account? Because obviously customers who are engaging with brands online, in person, you name it, care about that kind of thing.

- [Frank] Yeah, look, I think whether it's a blessing or a curse, business has had to deal with this already, right? GDPR, we've seen regulation in the U.S. now. So I would say a good chunk, probably if I had to give you a percent, 80% of the partners that we work with or talk with already have a data lake, right?

Appropriately structured, tokenized, right, and already protecting themselves against passing PII towards partners. And so I think largely the infrastructure as a whole already exists. And thank God we didn't have to reinvent that wheel or invent the wheel at all. But I would say that the cautionary tale would be, what's going to happen if OpenAI or one of these large language model providers or application providers changes their terms of service or privacy agreements. And so that, the scary thing about ChatGPT, in particular, experiences, you can share, essentially confidential information very easily and whether or not they utilize it, whether you've given them permission to utilize it, that should be something that major corporations are looking at.

That's why we actually prefer, right, to work hand in hand with those organizations to properly implement the technology and really walk them through the appropriate steps to say if we're going to share data, here is that process, and this is obviously what you should not be sharing with the interface or with the AI if you're trying to protect confidential information about your customer or about your business.

So that's at least the cliff notes on it from my perspective.

- [Neil] I think you're hitting on actually one of the biggest concerns I've seen in most businesses. It's not even so much about the privacy, and I'm not trying to disrespect that. The long term concern is what data will be allowed to use, and they look at much beyond privacy and that some of the rules, regulations, policy, legislation might change, and they're actually concerned about not just what data to put into the system, but what they might need as they develop the system, and if that changes in the future, does this mean they threw a bunch of money away?

Are you hearing any kind of pushback from any of the brands on that front?

- [Frank] Think back to November, December timing, right? When ChatGPT was launched, and it was Wild Wild West. We actually saw this huge influx of interest, adoption, and actually usage, and I would say in the last 60 days, you're seeing a pause from major corporations on, hey, let's go like actually create a center of excellence, right?

Or create a division now within this organization to determine best steps and best practices as we test out a new technology or bring on a new AI vendor. And so you're seeing that roller coaster ride where because this is so front and center in the mainstream media, you're immediately seeing regulatory bodies jump in.

You're seeing that, again, like I said, live on TV. This does give caution or pause to some organizations, but I would say overall, you're going to see adoption this year, and we're probably just at the tip of the iceberg. But yeah, I hope that answers your question. I think I need to go back to you a little bit.

Did I answer your question properly or were you going in a different direction?

- [Neil] No, no, I think, Frank, I think I'll call, you answered it partially, right? I think, I know that, again, that the brands, most organizations, that's, especially with their risk management and like in house counsel, are always concerned about the data we're allowed to use could change, it could shift.

This is just beyond even the privacy concerns. You think about something like data sovereignty, for example, if, people are certainly entitled to certain rights around their own data, you have to quote on quote license it directly from them, does that break a lot of the business models, processes already in place?

- [Frank] That's something again, that's going to take time regardless. The fact that we have legislation now coming into place that certain publishers will be paid by Facebook. We're talking, what is it, almost two decades after the launch of social media, and we're just now figuring out like licensing rights for content.

So you're going to see that same thing happen in the AI world. Now to what degree? I think that depends on the use case, right? And like where you're playing. I think you can't go wrong with playing with your own first party data as long as you are protecting, again, the privacy of your customer or whoever, again, you're sharing information from, but I would say as it comes to content, that one is, that's a messy world now.

You might know more than me on content, but the fact that people are sharing or getting links from ChatGPT right around content or even images and certain other AI applications, there's quite the road in terms of regulation and process. I think that lies ahead of us.

- [Neil] I definitely agree with that. It's been debated for, we've been debating with the UN for almost nine years now, but I'll say this, I know it's not very popular with most people when I say this. We talk a lot about the privacy aspect. I think we care more about data security. I honestly think privacy is dead.

Look at Millennials, Generation Z. They're very open because they're the social media generation. But, I always point out that, I was asked the question, have you seen the documentary, The Social Dilemma, and did you delete your social media after watching it? But we don't think about those things, right?

That's the thing. It's just the new norm or the next norm or whatever you want to call it. I'm not advocating to go public, put everyone's secret information go out there, but I think our concerns around privacy are overstated unless something bad happens. I think, I hate to say this, a lot of people have this kind of intuitive or intrinsic belief that they're taking my data and they're using it somehow, but somehow I'm getting some benefit from it.

So as long as that doesn't fall into the hands of bad actors, we're all good. That really seems to be the mentality.

- [Ryan] Sometimes it's the requirement to use something, for your benefit as a user, like terms and conditions, all those different things that we just blindly except and which actually made me think of the new season of Black Mirror just came out and the first episode is about kind of terms and conditions.

I don't know if you guys, yeah, so it's super interesting. Definitely recommend watching it just to see what could potentially happen with deep fakes and AI and simulation stuff, it's pretty cool. But it's, when I'm looking at stuff personally, and I grew up with social media for the most part, I started, I guess Facebook came out when I was in high school.

I want to use that tool, that platform, that thing, and in order to do it, I need to give permission, I need to give access, so that desire to use it is more of value to me than worrying about and, or being concerned about my access to or who I'm giving that information to. Now, I know I'm not, not everybody's this way, I know a lot of people are more cautious, they change all their settings, all their stuff, just to be as protective as possible, which I get.

But it's just really interesting how that has come up. And I think it's going to come up even more with these AI tools and how data is being collected, where it's being collected from and things like that.

- [Frank] You know, One point to that too, you're also talking about just the agility, right, of American businesses or businesses around the world. Facebook's original use case, right, was around education, school, and it was a way to keep up with classmates. You're going to see that same thing happen with AI.

And I feel like a company's mission statement and their use case is so important. So for us, we've stayed in our lane, for better or for worse, as it pertains to the company's growth. But we have a specific use case that we use that data for. Whereas some of these more general or horizontal models or applications, there's an endless amount of ways they can use your data to then monetize.

And that part is scary. And it should be something that is thought about from the beginning for a user. I do think the responsibility really falls on us to make those decisions. Because speaking from a regulatory perspective, it becomes very hard working completely handicapped business, which I think we know in a capitalist society, that's probably not going to happen.

- [Ryan] Yeah, have you also seen, I'm sure you have noticed, that a lot of the kind of more open and free platforms out there when it comes to people interacting have started to shut down access to or started to increase costs for getting access to their data through their APIs and things like that.

It's, starting to see it a bit more that I'm very curious to see how that starts to evolve. Because it goes back to your comment about who owns or not who owns, but how to monetize and how to drive value from content that's being shared or being put out publicly because a lot of times it's just been freely available, but some of these platforms and the places where this content is being published are now realizing, no, we need to start making money from this or we have something valuable here and changing the way they do business in that front.

- [Frank] Yeah, I think that's simply a byproduct too of the amount of market share they've captured. They're the, what, fastest platform to capture 100 million users in history compared to Netflix, etc. It's insane. So you captured that amount of market share up front, right?

And then the strategy or priority quickly becomes, okay, how do we monetize before people expect this to be free, right, for their entirety of their experience? And I'm with you that it does seem like almost it's a sudden jolt. But you're gonna see the big boys that are well capitalized continue to try to be free for as long as possible before the market forces them to do otherwise.

And then for players like us here, I think you have to be a little more like fiscally responsible, if you will. Our strategy from the front was we have to be very upfront with businesses on what these costs are and the services that we're providing to them. And sure, we have some free elements to our platform, but because of our mission statement and where we're heading as a business, we don't have necessarily the luxury to go monetize that data outside of our specific use case so.

- [Ryan] You bring up a good point. It really depends on who your intended audience is, who the users are going to be, who you're really trying to provide value to, to determine how fast, when, if you can start making money off of the things like that, like for instance, when we started IoT For All way back when, 2016, that's when Medium was free.

There was no cost to be part of Medium as free content publication, but then they started to either have pressure from investors or realized that we have all this attention, let's start to make, let's start to monetize it. And personally, my view is that they screwed up something that I thought was very awesome, especially for a publication like ours.

But there's still people who really love Medium and what it is with a paywall in front of it, with membership options, it's, you know, they really they change kind of the way they do things and that's definitely something that you're noticing or if you haven't, people haven't already, that you will start to notice more and more. But to your point, those who have the money and the backing can try to keep things free to grow their piece of the market and then figure out how to monetize maybe more native or different ways without detracting from the eyeballs.

- [Frank] Last point I'll make on this too, and I think it's fairly well documented now, but there's a great article that recently came out about just the sheer cost in training these large language models. Obviously, GPT-4, Bards of the world, Meta's probably unleashing theirs at some point, but we're going to see this kind of crossroads of these open source models, right, versus these kind of walled off garden models that we're already seeing. And that also allows businesses using open source models to be more efficient and more effective for businesses and users, in my opinion, than some of these larger language models trained behind the walls. Yeah, lots of expensive costs and initially building the technology for the large language model itself.

There's certainly ways to make them much more efficient, and I think between hardware and also again, just the way that these models perform and are optimized, we'll actually see some much more efficient use cases for businesses down the road.

- [Neil] So Frank, do you think then essentially crowdsourcing some of the training and model building is the wave of the future? There's no way, or there's no other path other than that?

- [Frank] Yeah, if I could predict that one, yeah, I don't know. I would say we'll see a combination of the two. I think regardless of whether or not it's open source or not, there's a variety of model training and unique value created behind the scenes on even open models. So, I am all for democratizing the technology.

In fact, I'd love that portion of what's happened with AI, at least initially. I have a feeling there will be some closed doors, right, and some competitive advantages between open source and walled garden. But I think also just based on what we've been able to produce, very efficiently, by the way, in terms of capital, there's no reason that these open source models can't be used to generate enormous value for businesses with greater efficiency than what we're seeing in some of the star powered names.

- [Neil] It makes a lot of sense. I call it a hybrid approach, work with DALL·E 2, with ChatGPT, that you do some crowdsourcing, it gets mundane enough or broad enough, I'll call it the David Lynch theorem or something, that there's enough use that people can go in and look, we all know the stats, ChatGPT, 10,000 users, 200 million users in four weeks.

That kind of growth is just unheard of, but I think it shows the power and diversity of what they were able to crowdsource. But now you have people, their own instances, paying for their own instances, doing their own private work and retaining their models, so it's a, it's an interesting dynamic because, I know Ryan and Nikolai have heard this from me before, generative AI isn't like new, it's been around for nine years, right? It's just the use of it was so niche, right? It was so very specific. It was not that broad appeal like we have today.

- [Frank] I still remember my first summarization model, right? My Q&A model. So yeah, one more point to you, Neil, I'll keep it short. I think also you see like diminishing business returns, right? So as we fought for accuracy in our model, at a certain point, you go, okay, what's the difference between 90% accuracy and 95% and are we actually adding that much more value to businesses?

Are we just adding to our costs and their costs? And so I think there's a world in which, like you said, there's just enough business value added at a certain accuracy, right, or capability level where going that extra mile sounds sexy on paper or for the media, but it doesn't actually mean anything, right, for the value we're actually generating for the world.

- [Neil] A hundred percent. That's the challenge, right? How much trust, what comes with what we need to build sufficient trust, not ultimate trust, but sufficient trust.

- [Ryan] Since ChatGPT has come out, how have you seen engagement, demand for AI tools from your perspective with the customers that you've worked with, interacted with? Has it raised more public concern about AI where people are now maybe a little more hesitant to adopt than they were before? Or is it more exciting and interesting and bring awareness to something that maybe they weren't thinking about and they're really trying to now bring into their business to help them do one thing or another?

- [Frank] World old question, isn't it? The optimist versus the pessimist on which way this is going to help. So I don't think anything's changed from that level, right? The same people that were doubting or not trusting it before are probably in that middle ground now where they're trying to decide for themselves.

Whether or not this is gonna be the way of the future or whether or not it's gonna take their job, I think for the optimists, we've proven time and time again throughout history that technology will find a way, regardless if it's useful. I feel like we've already proved that. From a personal perspective with Native in our business, I'll give you a stat. We went from last year, finding it somewhat difficult to get large enterprises in kind of the later stages of the sales cycle, right, for us as a business on a B2B software as a service platform, to this year, we have about 65% of our inbound interest is from the largest enterprise segment in terms of business size and scale.

And so that literally was like flipping a light switch, right? We're talking to our board, we're talking to our investors and be like, man, we are just struggling to get this stuff over the five yard line with large enterprise. This year, it's more than half of our inbound interest is from large enterprise that are just like we should be using this and how do we use it?

And they're coming to us. If that's any indication, like I said, and this is just the tip of the iceberg, I feel like 2023 is going to be a blowout year for the industry, it's already has been in some ways depending on how you judge it or measure it. And then 2024 and 2025 are going to be even larger, right?

Because you're still having those corporations that have to plan and strategically integrate the technology. And I think it was McKinsey, someone released in just a recent report that 60% of the individuals they surveyed, right, at Fortune 500s said they're going to implement it in the next one to two years.

So that excites me because, again, we're already growing quite dramatically compared to 2021, 2022, which only means this is going to get better. So that's general sentiment of at least what we're seeing behind our doors. But I'm sure that's even more drastic for some of the blockbuster names out there too.

- [Nikolai] What you're describing reminds me, there was a story recently, consultants are going to benefit a lot from this boom in generative AI. All these companies need to implement all this stuff. There's all these regulations. You mentioned McKinsey. So all these consultants, yeah they will win in this scenario.

As it relates to ChatGPT, I think the big turning point was just having an easy to use user interface, and you had action and reaction occupying the same space. I can do something, I can give an input, and I get instant output, instant feedback. And this is what made it for the average person, it's like why the iPhone and like all these different apps are popular.

You do that with generative AI, and not surprising that it exploded like that.

- [Frank] It feels like you opened Pandora's box a bit, didn't it? To Neil's point, it's, man, this stuff has been there. It just wasn't easy to do it. We talk to venture capitalists all the time and one of the lines is always UI and UX can be, right, a unique value proposition for a business.

And I think that's becoming increasingly difficult because how many times in our lifetime do we actually see that hit where it's like, hey, I created ChatGPT, and this is the interface, and it's so simple that your grandma could use it, and you're right, that's really all it took to open eyes of everyday people who really even weren't familiar with technology. I'm a hundred percent in agreeance with you.

- [Ryan] We've talked about obviously a lot of different things here, but when we go back to digital twins and their use as it connects to the AI space, AI tools and so forth, where do you see this going? What do you see this enabling into the future, into the rest of 2023, 2024, and so forth, especially as it relates to the market research, understanding customers, understanding users better.

How do you see this evolving or where do you think it's going?

- [Frank] When I started this business, we had an underlying thesis around just better connecting consumers with the other end of the value chain. And some of the stats that really caught me off guard were like, there's an 85% fail rate for products brought to market, new products.

Like, why is there that much of a mismatch between consumers who essentially are telling us what they want and then businesses producing goods or services that just don't match that. So, I think that's one of the beautiful things that can happen with digital twins is better alignment within market more efficiently, therefore less waste, essentially more sustainable, if we even want to use that word anymore. The other piece is I really hope it empowers some inclusion. Neil would probably have some things to say about this, but there's going to be bias within these models. I think the alternative in some of the legacy methods of executing research, whether that be surveys or focus groups, are even more biased and less inclusive.

And I'll give you an idea, players like IPSOS, I love you, IPSOS, in case you listen to this, but your surveys include less than 10% of U.S. consumers. And that means you get a 90% hole, right? And who you are surveying or gathering opinions from. And unfortunately, for most people in this world, those are minority or, again, under surveyed groups like Latino community of the black community.

By creating digital twins, you can actually help fill that gap. In fact, our only requirement for creating a digital twin around a consumer and therefore providing a voice is whether or not they participate in online feedback. And on a monthly basis, 49% of Americans participate in online feedback.

Those are the types of good mission things I feel can be brought to the table. To Neil's point, as long as a bad actor doesn't come in and screw this up for us, it's another good opportunity for us to actually better the world with technology rather than put us in the same spot.

So those are where I view digital twins being extremely useful from just a mission perspective and efficiency and inclusion. Can't get much better than that.

- [Ryan] Neil, any last words from your side?

- [Neil] I think it's been a fascinating conversation. I think the audience probably knows we've just scratched the surface on a lot of these topics, but Frank, I thank you for sharing what you're actually doing with generative AI, digital twins, and helping us pioneer into mundane use.

- [Frank] Yeah, I'm sure we'll be talking again. I'm excited to find out what's going on at the UN.

- [Ryan] Frank, thank you so much for the time. Fantastic conversation. Neil and Nikolai, great question as always. But yeah, excited to get this out to our audience. And Frank, we'd love to have you back anytime to talk about kind of anything going on over at Native, over in the AI space in general, especially with how businesses and companies can benefit and bring these technologies in to help them be better. So thanks again for your time.

- [Frank] Thanks, guys. Nikolai, Ryan, pleasure.