So What? Theoretical AI Usage

Well, hey everyone, happy Thursday. Welcome to So What the Marketing Analytics and Insights Live Show. I'm Katie joined by Chris and John. Howdy, fellas. That's our up top today.

Uh this week we are talking about theoretical AI usage. So in theory, Chris, what are we talking about? We're specifically talking about the new paper that we published, the AI uh, you know, what jobs can AI do. Uh, which if you are at all in our sphere, uh, you have gotten multiple emails about. And this paper comes in two flavors.

The two flavors that comes in are a helpful, friendly uh summary, and then a here's a whole bunch of data academic paper. But what this is a paper about is fundamentally back in was it February or March, Anthropic released a paper about labor, about the the future of of uh work and labor. And that paper was very, very, very widely cited. So you actually pull it up. And Katie, we have shown this.

I've lost track of how many times we've shown this on various podcasts and blogs and stuff like that. But this was uh their paper, and they well, the way they put this together was they used a paper from 2023 that said, uh using the ONET SOC database from the US government, here's 18,000 tasks and a panel of human judges, plus at the time OpenAI GPT 4, uh which is now a dinosaur, uh, were asked, here's this task. Could a machine do it in half the time as a human? And that was the that's what creates this blue line on this chart. And then Anthropic mined millions of clawed conversations with their clawed thing and said, Do we see evidence in these different professions based on the context of the chat that people are using AI in this discipline?

And the red line is what they came up with. And so, Katie, you and I have been doing a ton of work recently um using our uh the TRIPS framework by Trust Insights, which is uh a five-part framework to say, is this a task well suited for AI based on time, repetitiveness, importance, pain, and sufficient data combined with the artificial analysis benchmarks for could AI does AI have the capabilities to do these things? Can it see, can it read, can it, you know, law reason things out? And that was the genesis of our version of the paper, which said, you know, what could AI uh plausibly do? And what we where we land on that, and I'll bring up our version of the chart, uh, which requires a lot of scrolling.

Um so while you're bringing that up, before we get into our version, I first want to talk about um look. Can we go back to the anthropics version for a second? I want to talk about the methodology a little bit. And so I've mentioned on a few, mentioned maybe a handful of times, or you know, every call that we have is that I have done true clinical trials and academic research, and I am like a super stickler for methodology. So when I hear that the theoretical, and it, you know, again, to their credit, they use the word theoretical, but this is highly problematic because people look at this and go, Oh my god, this is black and white, this is what this means.

So they said, What could AI do that humans currently do? So that's blue, and then they mind chat conversations, probably without our consent. Uh, that's neither here nor there, that's for a different episode, of what of how people were actually using it. We know from our own, you know, talking with people that the limitation of people using these tools is they don't know what is possible. So I feel like that red section of how people are using it is the wrong way to go about it because most people a lot, you know, I'm saying not in the marketing industry or data science, but I'm saying the general population is still using a large language model as a search engine or write this email for me, or take these notes, like very basic functionality because they don't know what's possible.

So I feel like first and foremost, just acknowledging that this specific analysis is riddled with methodology problems if you're looking at it from a truly is this evidence-based, and I would say it's not. Yeah, and it's it is based uh one of the things about Claude in particular, Claude does not have a lot of the modalities that other tools do. It cannot generate images, for example, right? It cannot generate video. It is a very capable text-based tool.

And it does have a mobile app and it does, and the mobile app does have uh the ability to interpret images and things like that, and it's pretty good. But there are other tools, you know, chat GPT, for example, Gemini, etc., that have a lot more multimodality. And so in some of these fields, to your point, Katie, there are capabilities of AI that it can do that people don't know it can do. Or uh, as was pointed out uh many times on LinkedIn by folks like uh Ethan Mollick over at Wharton, people are using the free versions of tools in the general population that are so reduced in capabilities compared to the paid versions that they never truly experience what you know a fable five model can do because it's behind a paywall. So a classic example installation and repair here.

I use AI all the time for this. Uh, one of the top things I do in one of my keynotes is I show a video of me taking my phone and walking up to my furnace system, uh, turning on the camera saying, here's the model of my furnace, here's what it is, and I point to the valve and say, Why is this leaking? And it says, that's a pressure release valve. It's fine, just hand tighten it until it's tight and it will stop leaking. It just got loose.

And I'm like, okay, that saved me 175 dollars from calling out the plumber for what is clearly just illusory. I did Google it just to make sure that was actually what that part was. But that is a use of AI in this case that A goes beyond chat, right? Because it's multimedia, and B is an installation repair test where Anthropic reports no usage at all. Well, and I, you know, I feel like we could sort of debate the finer points of is that really just a Google search, or is it actual true use of AI?

But your point is made. Like your point is well taken where, you know, there's, and this is something that we point out in our paper is that a lot of these industries where there is no notion of can AI do it, there is actual opportunity, but it's not the AI is going to do this for me. So AI is not going to hold, you know, the torque wrench in a mechanic shop and like fix the engine, but it is going to guide you through perhaps maybe a different model of engine you've never seen before. And so it's thinking about it in terms of those assistant, you know, categories, those educational categories, versus it's going to pick up the tools and fix the engine for me. Maybe in some, you know, factories they have those, you know, robotic, you know, assembly lines.

That's one thing, but it doesn't extend to your everyday mechanic. And so I think that that again is sort of the some of the blind spots in this analysis. Totally. And we are already seeing advanced use of AI in agriculture, for example. Uh, there are robots that are autonomous robots that have been deployed to fields that have computer vision and essentially look at a tomato plant, for example, and assess based on leaf coloration and things, does the do we need to make an adjustment in the fertilizer mix or whatever?

And it does it fully autonomously. There's no human chatting with Claude about whether this tomato plant needs maintenance or not. It it just is this gigantic machine that moves through the field. So there are, as you said, Kate, there are a lot of gaps in this because you're not using an LLM to do that work, but you are 100% using AI because you're using computer vision combined with the vision language model to make that assessment. Go, yeah, this tomato, the soil here needs more calcium.

The leaves are turning this color because there's a deficiency of calcium, and then it looks in the eight tanks of chemicals it has and says, okay, let's spray some fertilizer on this. What's also not acknowledged here, because you just came up, you just brought up a really good point, you know. So you have these like robotic systems. Um, it doesn't speak to the availability of these resources. And so you mentioned like things like Fable Five, they're behind like a paywall, you know, and so if it's if Claude was just scraping free chat conversations, then yeah, I would fully expect those conversations to be fairly basic.

Whereas what you're describing, someone has to have access to some sort of like autonomous agentic robot that is, you know, is pre-programmed to go out into a field to be able to, you know, assess soil and temperature and you know, other things, that's not captured in this analysis. So when someone is looking just like, you know, at this graph, whatever you want to call this, it doesn't tell that story, but it does scare the bejesus out of people. Yes, that is true. So our version of this uses job descriptions instead, because in a lot of cases, there are things that are in job descriptions that's uh to you uh to your point, Katie, uh, multiple times, someone may not think, oh, I could use an AI tool for that and let me try this out. And and they see the results from it and they go, Oh, I didn't know it could do that.

So it's one of my least favorite terms, but it applies here. It is the art of the possible of saying, like, what can AI do? And so what we did was we fed 90,000 tasks through two different models, Claude Haiku and uh Deep Seek, no, it's three different models deep seek before and minimax M3, and ran through the TRIPS framework plus our AI likelihood uh index to say of this job description what the tangible outputs of this job, and then of those tangible outputs, what is the likelihood that AI was would be capable of doing those tasks? And what we end up with, which is our red line here, in a lot of cases, there are big chunks of jobs, even in things like you know, computers and management stuff where yeah, like talking to your team, coaching people, collaboration, machines are not going to do that, or you probably don't want them doing that. And then in other professions like production, buildings and grounds stuff, healthcare support, there are jobs that there are tasks that machines can do.

People may not be having conversations about them with Claude right now, but they absolutely are in that wheelhouse. And so that's where we at landed with with our version of this chart. Well, and again, it comes down to availability. So our good buddy Brian, hey Brian. Uh, he said, when we start seeing AI being more integrated into robots, these blue collar gaps are going to fill quickly.

I agree with you to an extent. I think the theoretical, those blue lines are going to fill in more quickly, but the red, the actual, can we do it? Do we have access? Uh, I think that's gonna be tough because it's gonna cost money. It's an expense that people may or may not be willing to take on.

And as we know, AI or robot or a machine, whatever it is, is not a set it and forget it. It needs maintenance. You need someone on your team who actually knows. Like, let's say the robot, you know, breaks down in the middle of the tomato field. Like, is anybody gonna fix it?

Does anybody know how to fix it? Is there a reboot button? Um very quick side note. It reminds me of one of my favorite cult movies, Return to Oz. There's a character called TikTok, prior to the actual app.

Um, and he breaks down, and there's instructions written on him about how to get him restarted. But without those instructions, like this, you know, smaller version of Dorothy is just standing around going, what the heck do I do? I have this robot, this mechanical man standing in front of me, and I have no idea how to boot him back up. And I feel like that's gonna be the challenge as you bring these tools into less conventional industries. That's not to say people won't be able to figure it out, but it's gonna be a steeper learning curve because it's not part of their everyday use.

This is something that I have the privilege of hearing about from uh my husband who works at a company where the overlords believe very much in AI everything, but the everyday staff members, the team members, are actually like working with their hands, dealing with people. So they just keep having AI thrust upon them. Nobody has time for it. It's a whole different workflow, and there's nobody around to fix it when it breaks. So I just want to flag those things of yes, AI is going to become more capable.

No, humans are not going to want to use it more. It's my little there's actually a video from UbiTech that also shows uh kind of the reverse of that tin man thing. This is a total distraction here, but uh what the heck we'll go with it. Uh, this robot goes and it pulls its own battery out of its back uh and uh goes to the charging station, grabs a new battery for itself and and recharges itself on the floor, which uh is I think is is a lot of fun. So the the tin man uh being able to to repair itself.

John, what do you got for me? Help me. Well, it's gotta have two batteries, right? I mean, that's the uh the problem with that one. All right, so back to the point.

Yes. So back to the point. Uh there is a lot more that AI can do uh in those low represented fields, and there is less that AI maybe should do in the highly represented fields. So one of the fields that was most highly represented was business and finance. According to the Ilondo study from 2023, uh, 94% of the tasks in that field machines could do twice as fast as humans.

I I don't know about that. And the other, there's actually one other interesting blind spot here, too, that you mentioned, Kate that just occurred to me. Claude is a wealthy AI. Meaning you have to have money to use it. Like if you're using Claude Max 20, for example, that's a $200 a month subscription.

When you look at the universe of AI and what people are actually using uh out in the field, uh it is Chinese models. Uh Deep Seek V4 Flash, the new version of uh uh V4 Flash is an incredibly uh powerful model. It is very, very, very smart, and it is one fortieth the cost of uh Claude. Like it is literally uh a penny on the dollar uh for Claude. If we look at open router's most recent leaderboard, that deep seek for uh V4 Flash is uh six trillion tokens.

It is uh the largest model in use uh in in the open router development system, followed by high three by Tencent, MIMO, the new version of Flash, GPT 5.6 Luna, the smallest version of OpenAI, Claude doesn't even make it in the top 10 because of its cost. And so a paper that is based on Claude usage inherently is like, well, it's not even the top 10 of the most used models in some of these things. So how do you how can we know that's even representative of what the AI universe is? Well, and I feel like also one of the distinctions you just made, and you know, maybe I'm mishearing, but you're saying in the development space, and so if I'm in healthcare and I'm, you know, a nurse and I'm being told to use AI, my first thought isn't, huh? I wonder if there's a cheaper Chinese model that I can use.

It's what is right in front of me. So I I hear you when you say that these other models, these alternative models to like the big three are being used more. But what this paper doesn't say is by who? Who is using these things? What are their professions?

What is their level of expertise? Because again, if it's pulling, and I'm not and uh I'm thinking of the Claude research specifically, not ours, it's just pulling chats. It's pulling chats from literally everyone. So it could be your 83-year-old grandmother who's looking for you know a blueberry buckle recipe because she just got a bunch of blueberries from the you know farm stand. That's represented here, but she doesn't have a profession that's represented.

So it's it's definitely skewed and you know, misrepresented data. So I just I also want to sort of acknowledge because when the general population sees data like this, we tend to panic because the AI taking my job panic is really real. And these papers don't do a great job of really explaining what it means. So hopefully that's what we're able to do today, in addition to providing our own version of the research. Exactly.

So that's that's the short version of what we found that there is much more opportunity in the underrepresented fields. Um, and we were talking about this on the podcast this week. You know, there's if you are a sniper um and you're and you're doing counter sniper operations, you can use vision language models to assist you. If you are a lifeguard, uh you can use a vision language model to assist you. That's not gonna be represented in this because probably most people are either don't know the capability exists or don't have the knowledge for even how to ask the machine for help with something like that or to build a system that is agentic in nature that could handle a task like that on an ongoing basis.

Okay. So the other thing that was interesting that is that the anthropic numbers, when you look at them in terms of how well do they represent the the goodness of fit to all the different categories, anthropic clawed usage numbers only have a goodness of fit of about 33%. And what that means is compared to what's possible, the way people are using Claude is not a good fit. It it inherently says the way people use Claude is not a good fit to the way AI could be used. And uh our uh score is a much closer fit to that to say, like, yeah, this is this is very likely a much better way to say, yeah, if you're gonna decay if you're gonna try to figure out how should we be using AI, uh, the methodology using the TRIPS framework for trust and size is going to get you better results.

Which I think is more helpful for someone trying to understand where does AI fit in to my business? Because that's really what we're talking about, is you know, anyone who you know from any walk of life is welcome to read this paper, but our ideal customer for this paper is someone in a business who is trying to figure out that next step in their AI enablement. Um, and so for us to ground our analysis in that makes a lot more sense than I think, you know, just taking a look at uh what are people talking about? What questions are people asking these chatbots? Like, that's not helpful.

Yep. And it's interesting because when you start digging into the individual roles themselves out of the 90,000 different tasks. This one is food and protein production at Cargill, uh, you can start to see what are those tasks that are uh either lower or higher value. So recording production and inventory data and computer systems, yeah, that is clearly an AI task. Whereas down here, manual meat and protein processing.

I I think your husband would agree, yeah, probably don't want AI doing that. No. Uh for those who don't know, my husband's a butcher. Um but no, I mean, and and that's exactly it. Like this passes what you call the sniff test of logically that makes sense.

If you really dig in, there's probably ways to incorporate AI, but at broad strokes, that that all tracks that makes sense. Yeah. And that's and to your point, that is how we validated that the roll-up results were good is by looking at some of these 90,000 tasks through the uh, I think it was 3,600 jobs that we dug through to see are the are the tasks that are flagged as good for AI actually good for AI, or did something go wrong? Like if it said, yeah, shopping meat is a good task for AI, that would be an indicator that something went terribly wrong. Mm-hmm.

So, Katie, based on the papers, based on the analysis, based on all these things. If someone's saying, okay, I understand the theoretical usage of AI now, how do I how do I make more uses? Our company has spent a gazillion dollars. We've budgeted 11% of our operating expenses next year, and we don't know where to start. How should somebody get started with this uh with this sort of AI enablement?

I am so glad you asked. Uh, we can help. Trust insights.ai slash AI enablement. Um which is truly we really can help you. But in all seriousness, I think that what we've seen a lot of our clients doing successfully.

So it's not that people are uneducated about AI. It's not that people don't know where they want to take it, it's that oftentimes they have limited resources or the shape of their organization is the limiting factor. In the Inbox Insights newsletter this week, which you can get at Trustinsights.ai/slash newsletter, I talk about the four basic structures of an organization and the limitations that it puts in place if you're trying to scale AI. So most enterprise-sized companies, and I'm saying most in broad strokes, are a hierarchical organization where basically, you know, if John is the CMO and Chris is the VP and I'm the analyst, I can't talk to John. I have to talk to Chris, who can then talk to John.

And so those layers moving up and down provide a huge block for AI to be able to scale. Then you have functional organizations where you are basically clustered by disciplines. So you have marketing, HR, blah, blah, blah. You can scale AI really well within that specific discipline, but then it doesn't translate if you know finance is doing similar things to marketing, but the two organizations are sort of, you know, that virtual firewall. That's another different limitation.

You have a flat organization where people have a lot of autonomy, but really nobody's in charge and nobody has authority. So nobody can say this is how things should be moving forward because nobody is used to having to answer to anybody. And then there is a matrix organization, which on paper is the most ideal for scaling AI, is the least ideal in a business operation sense because it brings resource and billing headache. So a matrix organization, let's say I'm the manager and Chris and John are my two direct reports, but then I also have another team under me, they can all share resources, but then Chris is overbooked, and then John doesn't have enough. And then, you know, who's footing the bill and who's making that decision on resources?

It's there's more of a collaboration for AI, but less uh more of a headache for the humans. And so these are the limitations that AI runs into, but it doesn't know that that's the limitation. And so this is where a lot of organizations are today. So first and foremost, I recommend people start with getting a sense of what those blocks are. And I would start with looking at your organizational structure because once you understand what the blocks are, then you can start to like pick apart and prioritize how to move them.

So if you're in that hierarchical organization where only certain levels can talk to certain levels, you know, it may be an opportunity to revisit your AI champions program or your center of excellence to say, is it doing what we need it to do? Is it pulling people from across these different levels so that they can actually collaborate together? And then so on so forth. Um I would take a look at your use cases. This is something that a lot of organizations they have.

They're like, here's all the things I want to be able to do with AI, but I'm not currently doing it. And then you can run it through what Chris mentioned is the TRIPS framework. So we have a free version of the TRIPS framework on our website. You can go to trustinsights.ai/slash trips, and you can do a very simple exercise where you're scoring each of these tasks for time, repetitiveness, importance, pain, and sufficient data. And that will start to give you a sense of priority of where to move next.

And then if you really have a bigger scale project, that's where we come in and we can do a lot of this for you because we have a lot of this automated with our own proprietary software that we can do bigger scale things. That's all right. The other thing I would recommend is from a technological perspective, it is good to check in like once a quarter to see what the capabilities of the tools themselves are and understand what is out there. Um, because uh, you know, and one of the sites that I spend probably the most time on is one called artificial analysis. It's a company that does benchmarking, independent benchmarking of all the different uh AI models and things like that.

And you can get a sense of essentially who are neighbors, right? So Claude Opus V and Claude Fable, they are neighbors. They're very close in terms of intelligence. Uh, when you look at their agentic capabilities, you know, they're they're somewhat neighborly. And there's a spectrum of all the different companies, 594 major model families out there.

But the on the leaderboards are other things like text to image. You know, what models are are good at that? What models are good at image editing, what models are good at text to speech, which ones are good at transcription, speech to text, and so on and so forth. And as you're starting to think through, especially when we're in budgeting season, it's a good idea to take that step back and say, what are our use cases? What are our needs, as Katie said?

Um, do the TRIPS framework and then say, do is there an ecosystem mismatch? For example, if you are a Microsoft shopper and you're using Microsoft Copilot, and Copilot's primary model is Microsoft V4 uh multimodal instruct. When you look at V4 on these charts, you go, oh well. Um in fact, I think I can even pull up uh Microsoft V4 multimodal. This is the model that powers co-pilot.

Look where it stands down here on the intelligence indexing question. Well, um, and our friend Erica brings up, you know, as we were talking about organizational structure, matrix organizations can also end up with every marketing team. And this is when you get into the function digital corpcoms content using different tools for different functions. Um, that's exactly right. And so this is why I recommend starting with what kind of organizational structure do you have?

Because you could have on paper a matrix organization, but then within that, then you have a functional organization, and that's where you start to run into those issues again. Once you know what that looks like, I think the next step there, to Erica's point, is to do a tech stack audit of what are you using? Why are you using it? How are you using it? And then this leads you into your use cases of like what is it that we actually want to do?

It's a lot of work, but if you are an organization that is looking to really scale AI in 2027, this is the work you have to do. Exactly. And it is occasionally useful, uh probably more than a case, but occasionally useful to bring in that outside perspective because a lot of the time, even in in any organization, this is true of all organizations, even as one of the smallest trust insights, you kind of get used to doing things a certain way. You have a certain tool set you're working with and things. And if you're not ever getting out, uh, you know, ever leaving your desk and getting out in the world and seeing what's out there, you may not see some of those use cases and go, oh, I didn't know it could do that.

That is very, very different. It's one of the reasons why uh I spend a lot of time reading uh academic papers, published papers from the major AI conferences to see like, oh, this is you know, this is how they're doing this in drug discovery. Like, I didn't even know think about using a language model in that capacity, and now I have that, you know, mentally I can add that to my tool belt of things of things to try. The way that I describe that in this week's newsletter is sort of like those blind spots. If you think about the layout of your house, there's everyone has issues with the layout of their house, regardless of the house they live in.

But we just get so used to, you know, working around it because it's easier than taking the time and resources to change the layout of your house. So like if your laundry is in the basement, but it would be better on the first floor, you might complain about it, but the time and energy and budget that it takes to fix it, you're just going to work around it. You're just going to figure out a different system to make it work. I mean, John, you love slash loathe home repair. And I'm sure you run into this a lot of like if it was just moved like this far over, or if we had just thought to, you know, swap these rooms, but you learn to live with it.

Yeah, carrying a washing machine up cellar stairs is uh it's cheaper to just leave the trash in your basement and buy a new one, you know. It's uh switching costs are real and yeah, and design is critical, right? Like having it done. Um, you know, if the shutoff to your oven is behind it and you can't move the oven, you're just gonna burn the house down. You know, you need to you need to think about this stuff and take it into account.

The last thing I'll add on theoretical AI usage is, and this is something that was from two weeks ago in the newsletter you wrote, Katie. Um if you have a something, you're not sure if it's a use case, ask the machine. Hey, how could you do this? If you, you know, if if I give you this task, how could you uh tackle it? Or what parts of this task could you do?

Like, yeah, clearly you, you know, the machine is not gonna be handling, you know, cutting meat for you. Uh however, it might be able to teach you a thing or two about you know recognizing different patterns or things to look for, or hey, that's got cyclospora all over it. I don't even know where to go from that, but uh I think basically the use case that you're describing, a lot of it is educational. Um, I would say that with the caveat of one of the things that we know about these large language models, it is that they are very over-eager and they are programmed to be incredibly helpful and say yes to everything. And so if you give a large language model a use case without any real guardrails or the you know restrictions that you have or the things that you and your company can't do, then you may be sort of led down a false path to be like, well, AI said it could do it, so let's just have it do it, and then you find out along the way you have all of these different bottlenecks, or you know, you don't have the capability.

So I would say ask AI, can it do it? But also say, and here are the limitations, and here are the things that we as an organization can't do because of budget, because of resources, because of education, whatever it is. Can you work within these limitations? Yes or no. That's gonna give you a better quality answer.

Yep. Another thing, and you you do this a lot, Katie with Claude Cowork, is to go into add a connector to you, whatever your company's collaboration environment is to say, hey, pull out the last seven days worth of tasks that I've talked about in our system and just make a catalog of it. And then you put that through the TRIPS framework to say, okay, well, what did I do this week? Oh, I did all this stuff, and I probably don't need to do half of it because a machine should be doing like this, like, oh, look at the yet another status update. Like, yeah, you should you should have handed that off to Claude a while ago.

But there are a lot of the time we have task blindness. We get so used to doing something that we we don't even physically remember it anymore. Like, oh yeah, did I do the annual report? Did I do the monthly report? I don't remember if I did the monthly report.

I don't remember what's in it. That was before AI. Um, if you if you do use these tools to dig into the knowledge bases you have and audit the systems and audit yourself, look at it, have it look at your outbox, have it look at your your direct messages and say, What did I do this week? I do do that a lot because I can't for the life of me remember anymore for a variety of reasons. But I found it to be a really incredibly useful tool because then I can push it into our project management system.

And then I am not, you know, solely responsible for remembering everything that I committed to. Um, you know, my next evolution, it sounds really basic, is to start using AI a little bit more in my inbox because right now I kind of keep everything in there. I'm a bit of a hoarder in my inbox, and it gets overwhelming, and there's stuff in there that I don't need to be keeping in there. I can move it into more useful places, like a task list, or I can archive things that I'm just CC'd on that I actually have no action to take. Like there's better ways to do it, and it's very much as you said, it's sort of that task blindness of like, well, this is how I've always managed my inbox.

Why would I do it any differently? But it's getting overwhelming. And to Brian's point earlier, as you start getting the systems, and this will be happening over the next year or two, as you start getting the system integrating into the physical world, you're going to see many, many, many more opportunities. Uh, for example, uh, if you've ever seen a little Raspberry Pi, the little tiny, you know, tiny super cheap computers, you can get those with tiny little cameras. And if you wanted to, I've done this myself.

Uh, you can plug that little computer into your laptop or whatever. Claude or the tool of your choice, chat GPT, copilot, whatever, can write the code that those little boxes speak, because they're really just Linux boxes, and you can start building your own real world things. So, for example, uh, if you can take a Raspberry Pi, a little camera module, and have a tool build a bird watching app that you know you put next to your bird feeder, and when a bird lands, it turns the camera on and starts sending you screenshots of here's what this, you know, here's what bird landed on your feeder, or build your own flock camera or whatever. Um, but if you have things like Arduinos, uh, which are robotic systems that you can plug, you know, a little connectors to, you can start to build in the real world using what AI using AI to engineer the software for it. So now, in terms of theoretical use of AI, we're getting out of the black mirror and getting into real life.

Um, there are a number of, for example, open source Wi-Fi systems. If you want to construct your own Wi-Fi network at home, easy to do. We do this a ton with servers that we operate, like our email marketing automation server. I have Claude Code on that just helping me maintain the sucker. If you have, if you go on Amazon or the you know, device uh vendor of your choice, there are weather stations that you can buy for your backyard that have open software, so you can modify them to your heart's content to say, like, yeah, notify me when the backyard you know temperature gets above this, or when it's pouring rain or whatever.

Um as realize some of these use cases are can't you just look out the window? You can look out the window, but you don't necessarily remember to record the data. Um right now, for example, we are entering what is called a super El Nino, uh, where the weather patterns in the Pacific actually switch directions, and you're gonna see map very large climate changes across the planet because the effect of the Pacific Gulf stream is reversed. So if you have that data plus an AI tool, plus your own backyard data, you can start to project and and measure its impact and go, okay, if I'm growing crops, as we saw on the food production and farming section of the theoretical usage. Now you have systems that can help you anticipate and get ahead of problems before they cost you your entire season.

I mean, you had me at you know, bird watching app. Don't get me anything for Christmas, build me a bird watching app. John, what do you think? Uh, you know, you're on the front lines of hearing how people are, you know, trying to scale their AI and what's next for their AI enablement. Like, what are your thoughts on like where people should go?

Yeah, well, the a key point to this is you know, theoretical is great to discuss, but the reality is there's human factors that have to be part of it, right? I mean, theoretically, there's no reason we shouldn't all already have electric electric cars, right? I mean, it's technically superior in every way, really. Um, so you've got to align the human forces, you've got to get incentives square so that there's, you know, because we know that we know orgs right now where the cultural and political issues will actively fight against this. And you know, not only even though it's profit motivated, it can still be sabotaged.

So yeah, you know, you need to uh get ahead of it, you know, get people aligned and trained so that they can understand the opportunities. And uh yeah, I mean and we're of course glad to help with this kind of stuff. You know, we've worked in other orgs that have struggled to do this stuff. It's not enough to just check boxes and say, okay, we're using it, you know, and it's well, you know, there's more to do than create bad pictures for your slide deck. I mean there's other places to go.

So um yeah, you know, take full advantage. There's there's never been a greater opportunity as far as business, you know, in the last 20 years, you know, since the dawn of the internet, this is the next opportunity to get ahead of everybody else and find a new way of doing business. So uh take your shot. Yeah. I have I'll give you one last example.

I was doing this for a self-storage company that has uh gate software hardware and software that's basically controls like the front door and the you know the lockers and stuff like that. Uh and this one company uh was going out of business. No, there's one company saying we're ending our desktop software package you now all have to pay like a hundred X more for the SaaS software version of this that's internet control. Out of curiosity, I said okay, can we find the hardware specs for this system? Yes.

Can you know and it's all made of very common components so on a LARC I had an AI system just build net new software for it that because it's not a complicated system to begin with. Um and produced it and you know it it passed all of its uses I don't own a self storage facility so I had no way of actually testing this in production. But it'll if you have hardware today that has an open standard. There is no limit to what you can do with this. So the theoretical AI usage just goes way beyond to summarize this email into hey.

I just saw this company that produces you know open source hardware for like you know wearables, like you know, watches and things. I'm gonna make my own fitness software. What the heck? Why not? Um let I would say let folks, you know, folks should figure out okay, what is what could you do if you have access to the system and you can and you can get a machine to talk to it.

What could you do? And and I guess we're we're making a bird watching app. Yeah. Um, well, I was gonna say we have so we mentioned two papers, our academic paper and then our more user-friendly paper. We have links to those papers in our free Slack community, Analytics for Marketers, which you can join at trustinsights.ai slash analytics for marketers.

Um, if you want to learn more about you know how we can help, you can go to trustinsights.ai slash contact, talk to a really human. He's right here. His name is John. Um, or you can learn more specifically about how we can help you with your AI enablement. Go to trustinsights.ai slash AI enablement.

And then you don't have to talk to John. But if you would like to, you still can. It's fine talking to John. Not Ricard. You don't have to, only if you want to.

I love talking to John. I don't know why you wouldn't want to. All right, folks, that's gonna do it for this week's show. Thanks for tuning in, and we will talk to you all on the next one. Thanks for watching today.

Be sure to subscribe to our show wherever you're watching it. For more resources and to learn more, check out the Trust Insights podcast at TrustInsights.ai slash TI podcast at our weekly email newsletter at trustinsights.ai/slash newsletter. Got questions about what you saw in today's episode? Join our free Analytics for Marketers Slack group at trustinsights.ai/slash analytics for marketers. See you next time.