From ChatGPT to Agentic AI: The Good, Bad, and Ugly of Corporate Automation

Speaker A: Mean, have you seen what Shopify CEO? I mean, you know, we've talked about CEOs in the, in, in couple of episodes back and you know, AI replacing CEOs, it looks like someone is taking notes and trying to jump ahead and replace their employees with AI.

Speaker B: I did see the memo you, you're talking about from the Shopify CEO. Only because you shared it with me. I, I was, uh, thankfully blissfully ignorant of that one, uh, until you shared it with me. But yeah, it's pretty interesting. Uh, I think the kind of official line is no new hires unless you can make the case that AI cannot do the task for you. Is that, is that a pretty good, uh, summary of the, of the, the key takeaway?

Speaker A: Yeah, I mean, some of it was good. Some of it was just like.

Speaker B: Yeah, I mean, it's interesting, right? I mean, if you think about companies at the C suite level, um, whether it be a CEO, a CFO, uh, probably increasingly CMOs like re reducing the cost to deliver, increasing efficiency, um, you know, finding ways especially to reduce things like your indirect labor, right? This, the labor that's not related to delivery of services, anything that you can do around, um, process automation, um, and then even. And then of course on the delivery side, you make your business model even more scalable to the extent that you can use, um, you know, whether it's simple process automation or more advanced kind of agentic type, you know, AI bots and things like, I mean, honestly, I, I sort of get the sort of bean counter financial numbers around. Why it's, why it's appealing, I think where. I think where the techno pessimists, uh, like myself need, uh, to push back is just kind of finding the guardrails and the limits, right? Like, how far can you push those things before you start to lose out on things that are really valuable, that actual reasoning, thinking human beings bring to the table? And I guess the next few years are gonna be really interesting finding out where that balance point is, right?

Speaker A: Yeah, I mean, we've been using AI like our clients have been using AI. We've been using AI. Our partners were using AI.

Speaker B: Hello and welcome to the Digital Disruption Podcast, the show that takes you on a journey through the ever evolving world of, uh, digital transformation. In each episode, we explore the latest trends, strategies and insights that are reshaping industries and driving innovation. Whether you're a business leader, technologist, or simply curious about the future, this is the podcast for you. Join us as we deep dive into the transformative power of technology and its impact on businesses. Society and our daily lives. Get ready to unlock the secrets to success in the digital age. Let's begin.

Speaker A: Hey, good morning, Patrick. It's morning, right?

Speaker B: Almost afternoon, but it's still morning. Dawn. Good evening to you.

Speaker A: I called the morning because we're always having our meetings and just podcasts in the early morning on. On your clock, so.

Speaker B: That's right.

Speaker A: Oh, how. How. How's everything been going?

Speaker B: Good. It's finally turned into spring here in Billings, Montana. It's almost time to get the boat out, and, uh, I've been able to. To get up into the mountains a little bit and, uh, do a little bit of shooting here and there, a little target shooting. Uh, got to get my bow back out, get some archery in, so. Been having a good time. I was in Cairo and Istanbul in February, as you know, um, which was a lot of fun. And, uh, we got to have some. Some coffees and, uh, and spent, uh, some time visiting, which was really nice to do in person. So.

Speaker A: Yeah, it was great having you on the summit and the rest of the company and, you know. Yeah, hunting is one of the things that hopefully, hopefully, AI, uh, will not cross into because it looks like I'm gonna get AI in my cereals at the morning. You know, I swear, I swear, I wonder what AI Tastes like. It probably does not taste good, but, you know, in everyone's mouth right now. I swear it's been shoved down our throats from everything. I swear to God. I mean, have you seen what Shopify CEO. I mean, you know, we've talked about, uh, CEOs in the. In a couple of episodes back, and, you know, AI Replacing cells. It looks like someone is taking notes and trying to jump ahead and replace their employees with AI.

Speaker B: I did see the memo you. You're talking about from M. The Shopify CEO. Only because you shared it with me. I. I was, uh, thankfully blissfully ignorant of that one, uh, until you shared it with me. But, yeah, it's pretty interesting. Uh, I think the kind of official line is no new hires unless you can make the case that AI cannot do the task for you. Is that. Is that a pretty good, uh, summary of the. Of the. The key takeaway?

Speaker A: Yeah, I mean, some of it was good. Some of it was just shit, like.

Speaker B: Yeah, I mean, it's interesting, right? I mean, if you think about companies at the C suite level, um, whether it be a CEO, CFO, uh, probably increasingly CMOs, like reducing the cost to deliver, increasing efficiency, um, you know, finding ways especially to reduce things like your indirect labor. Right? This, the labor that's not related to delivery of services. Anything that you can do around, um, process automation, um, and then even. And then of course on the delivery side, you make your business model even more scalable to the extent that you can use, um, you know, whether it's simple process automation or more advanced kind of agentic type, you know, AI bots and things like, I mean, honestly I sort of get the sort of bean counter financial numbers around why it's appealing. I think where the techno pessimists, uh, like myself need uh, to push back is just kind of finding the guardrails and the limits. Right. Like how far can you push those things before you start to lose out on things that are really valuable that actual reasoning, thinking human beings bring to the table. And you know, I guess it'll be the next few years. It'll be really interesting finding out where that balance point is. Right?

Speaker A: Yeah. I mean we've been using AI and like our clients have been using AI, we've been using AI, our partners were using. AI in itself is not, it's, it's a tool. You can use it, you know, for the bad, you use it for good. But I think what pisses me most from that CEO like note is it disregards how much value a human can add to um, a role that he jumps right into. If we look at our podcast, let's say for example, I don't imagine us interviewing an AI because AI will not have the background experiences, the knowledge, you know, it hasn't been in the trenches trying multiple things and failing. Trying multiple things and winning. It doesn't have that, you know, touch of there is a human. There is a, there is a human side to the business. Even the business does. And corporate seam, you know, all dark. And it reminds me of severance. If you have not seen severance, you should definitely do. It's on Apple Plus.

Speaker B: Uh, oh, I did watch that. Yeah, there's some, it's pretty disturbing, um, series in some ways.

Speaker A: Yeah, um, like, like this is what like CEOs think that companies and like employees do. They just like, you know, and then, you know, we're done. But this is not, this is definitely not what, you know, I mean, yeah, AI is going to replace some low level, um, jobs. A lot of things can, can be automated. But if you try to hire someone, let's say for my role, which is, you know, creating content for essence. Yeah, I do use AI, but good luck trying to replace me with AI. Good luck to. Yeah, yeah, yeah.

Speaker B: You know, you bring up an interesting point. And um, I mean for me just watching things unfold, right.

Speaker A: We're not going back.

Speaker B: Like, like artificial intelligence has this large umbrella, right of different like methodologies, models, um, ways of you know, creating sort of human like reasoning outputs, uh, and task delivery. Right. There's a lot of interesting and powerful things happening.

Speaker A: Um,

Speaker B: the most interesting use cases are some combination of um, human intelligence design, you know, really designing uh, an appropriate use case solution and then applying the kind of computational power and scale um, and automate, you know, automation, automatability. Right. So there's as the things that I've seen that are most interesting are, are some combination of kind of art and SC to some degree. And I think that will continue to be the case. And you know, I think uh, in a lot of organizations, I mean it's like every conference you go to, right. Uh, every third piece of content on LinkedIn is like just like, like you said, it's just like forced down your throat this AI. Like it's just like it's either AI or if it's not AI, it's ML which you know ML is just like sort of one facet of AI, right. So yeah it's so, and just so we don't get stuck in our own algorithmic world or and and uh, acronym world. Um, you know, AI, artificial intelligence, ML, machine learning. So um, we uh, we should put the pause button sometimes and we should say like for example like you. And you also see all these post uh AI readiness. Is your organization ready for AI? Well let's, I guess let's kind of look at a different quadrants. Like, let's just look at like the big one. Uh, from a adoption uh standpoint, generative AI, right? All of the different GPTs or generative pre trained transformers like they're here, they're, they're. This isn't like are you ready for it? It's like you're either using them or you're not. And most people in a professional setting are using them at least for some purposes. Right. Like most people haven't like say totally transitioned from search engines to GPTs, but uh, but are using some combination or like are you um. Like some. One of the things I like to do is some um, basic data prep. Like if I pull uh, raw data poles from different data sources I'll use chat, uh GPT to do some basic data prep data normalization that would like. It's not good use of my time manually and I also don't want to write SQL to do it. It's just a one off thing for a very specific case. There's um, certainly a lot of the manual development of um, computer code that's kind of going away, right? So you have software uh, engineers that are you know, designing the, the framework and the schema of the solution. What should happen? Well, how the process should flow. But a lot of the kind of rote coding tasks are being taken over by, by GPTs. Um, so I think the idea that people ah, are still asking this question of AI readiness is in one way is like on one side of my mouth I'm going to say, well it doesn't really matter if you're ready, your people are already using these tools. Um, and honestly in all likelihood most companies have some licenses already for the enterprise or paid versions of some of these tools because you do get more functionality, more volume, all of these things. There's even like GPT aggregators, right? Like po, um, po.com, right? You can have a subscription and you can do queries on multiple GPTs. Um, and it's kind of charged like utilization of you know, tokens essentially. Um, so um, like that piece is already here, it's already being adopted. It's like, and you can even see the adoption curves on like chat GPT, like the number of new users that they acquired. It was like one of the steepest and fastest like acquisition curves ever, right, in the last like 40 years of any new tool, right? So um, and that ability, right to use basically natural language prompts, um, and to kind of start a quote unquote human like chat and to revise, actually revise that conversation, right, by refining the prompt. And then even like over time you yourself as a human, you get better at quote unquote engineering prompts. So you, you kind of, you get better at telling the GPT what you want it to do and the GPT gets better at providing you with results, right? So that, I mean that world like people should keep investing in, like there's no reason to think they're not going to um. But the idea that you know, like most companies or individuals are going to design one of these GPTs themselves, right? Like um, and M, more specifically the underlying large language models that those GPTs are built on top of, right? Other than like you can you like in GPT, like in chat GPT, if you have a license you can create a custom GPT, right, which has your own data sources incorporated APIs, etc. But that's not, you're not really designing a new, an entirely new GPT on top of the framework, right? You're leveraging the functionality that exists. Um, so most people, right, But a big piece of what's happening in ar, in AI are these large, you know, either foundational models or these large language models. That's like a lot of the adoption, um, by individual people and by companies is these, the functionality that's driven by those models. But most companies and most people are not going to invest in actually spinning up a new model. Right? It's profoundly expensive. Like, even though the costs have come down. Like it's, it's these, these models uh, have literally like um, billions of parameters, right, that, that all get like tuned and trained, um, over time. It's just not something. Even with the declining costs of computation, um, and GPUs is still not cost effective for your average, you know, company or individual to like spin up these models.

Speaker A: So.

Speaker B: But we're all going to take advantage of the ones that have been done and that will continue to be done. Um, we've already talked about generative AI. I mean how, how frequently, let's say in a day, in a week do you use some generative AI tool?

Speaker A: Oh, every day?

Speaker B: Yeah. How much every day?

Speaker A: I mean if I'm working, um, I

Speaker B: probably, if you're working, yeah.

Speaker A: I probably have Shajibity, Gemini. I, I don't use Bow. I use open router. I feel like it has much more like great models. Everything that, you know, comes up new, it's an open router. So it is great. I have a couple of uh, of like generative models out there. But I've, I've been loving Gemini. I um, use clothes, I use Shady because we as a company, I have a subscription, um, and I just, you know, I pretty much use it

Speaker B: two

Speaker A: to three times a day if I'm working on something, trying to find maybe more to be honest.

Speaker B: Yeah, I'm to the point. I uh, I'm like, I play with them because I, I need to for the work we do. But in terms of, for outputs, I'm probably to the point maybe once, once a day I use it for something. Um, I still like for search. I'm still kind of a search engine guy. I like, I still like Google for most, just sort of general search stuff. Um, so from that standpoint it's not a big, big contributor for me though. I know like we, in all of our clients, especially like our E commerce clients, we see increasing uh, amounts of GPT, uh, referrals in their analytics data. So from a consumer standpoint, it's definitely right here. You see a lot of the, the big SEO players are, you know, have already and are continuing to develop content about how to ensure that you're optim. Different GPTs. Right. Um, so, so it's important to kind of understand, like, what is your brand footprint, not only within search engines, uh, and within social media platforms, but now within these, you know, GPTs. Um, so these are all things every, I guess in that sense, like every brand and every person probably should be kind of playing in the generative AI, uh, space at this point to some, to some degree that, you know. Um, but how often do you, how often do you have a GPD output? Um, let's say like an email or a piece of content. And do you just copy and paste it and, and be done with it? Does that ever happen?

Speaker A: I mean, sometimes. Um, it depends on what I'm trying to do.

Speaker B: Yeah, I guess. For what purposes would you do that?

Speaker A: If it's something that I have not been exposed to before, I would definitely ask GPT to role play, um, the, like, the, the, the character that would output such, um, you know, such, um, such output. And then I'll let it, you know, use it, um, run it and just like give me an output that would get vetted by someone else. So that's what I mean.

Speaker B: That's what I'm saying. You would never just copy and paste that and like say send it to a client or publish it online, right?

Speaker A: No, it needs to be vetted.

Speaker B: Yeah. All right. I'm the same way. Like, I, to me, nothing. Like I just like fully automated communications or outputs and deliverables. I'm, I'm not there. I'm definitely not there.

Speaker A: No. I'll tell you a fun story. Um, so we get a lot of like, emails, cold outreach. We do some songs we're gonna hate on the people. Um, but I was getting um, some emails from someone. I don't know. It was a cold outreach. I wasn't interested. But there was like, like some kind of characters that. It's like an HTML, I think it was like bracket, like this bracket. And it has like pr, which is like break line. But I'm not sure if it was break line or not. And then the bracket was closed and it was all over the email. Like every sentence has it. And I was like, that does not look right. And uh, I got these emails, I got to the point like, uh, that does not look right. And I was playing with a lot of AI models and I think it was Gwen. So Gwen was The one who outputs these like characters. And uh, I was like, dang, dude. So you've been sending me fully automated emails without looking at them and like, like it's not the content that gave you out. It's. It's like that not refining the, the small bits that you know now like,

Speaker B: oh yeah, now I do. What I do think is really powerful is if you do the reverse. If you craft, if you craft really well designed, uh, fully vetted, you know, human generated original raw materials content like image assets, um, nice backgrounds, templates, um, and copy that's been fully, fully vetted. And then there are elements that you want to kind of like personalize and tweak at scale and in an automated fashion like based on inputs. Now that to me is really powerful and I know there's a lot, uh, on the creative side of digital advertising, a lot of personalization and, and like hyper. Hyperversioning of creative, you know, is happening. Um, efficiencies of workflows. Like, I mean, I think, you know, we've talked about this in the past, like the whole deep fake, you know, uh, versus, um, versus personal. I do think, I mean honestly, if it's, if you're, if it's literally like a uh, role of you know, like tape for a grip, you know, and you have 30 different colors. Like I think like using uh, a process to change the color of this roll of tape in my hand. Like, I don't think you're misleading anybody because you're selling, you're selling all 30 of those rolls of tape. Right? Like big deal. Right? But, but from an efficiency standpoint that's a huge win, right? If you can just update product colors. Right. And it's, you know, I get why people want to do that type of stuff. As long as you have the underlying assets are really good quality, I think that's awesome. That type of stuff is amazing. Um, but uh, so I, you know, I can appreciate kind of the scalability piece, the velocity at which you can create versioning, um, ideation. Like I know like creatives and design people, you know, kind of like doing like an initial ideation based on a concept and then you kind of, once you narrow it down to like three or four ideas, then you kind of start kind of using your more traditional design type processes. I mean there's a lot of interesting things people are doing. But like you said, excuse me, but like you said at the top of the episode is like, are like we need to push back on that sense of the full removal of the human I think there should be some noteworthy skepticism and cynicism around that, especially for some key aspects of not only what we do, but in any industry there's just a certain, uh, art and craft that is not negligible.

Speaker A: I think that eventually AI will develop such a framework like digital maturity that you'll have companies in buckets and each bucket is more advanced in AI. So you know, it's nascent, um, emerging, connected and then multi moment. You'll have also companies fall into like the same four categories where you know, it's minimalistically using AI to fully automating things like using AI and RBA and stuff like that and then everything in the middle. That's what I think because we've talked to a lot of people, um, with like leading brands and especially in mena, um, and we've asked them, how do you view AI? Do you think AI is going to, you know, um, make a difference and impact on your business? And most of them were, you know, we're seeing where it goes and we're seeing how we can incorporate it into our business, but it's not really making such a big, big difference right now. And mostly like most of the people that will talked with E Commerce and um, other brands like this. Um, but I mean, outside of the companies that use AI exclusively for their product, um, I think that, you know, companies will have to find a way to navigate AI because you know, as a company you can't say like, yeah, using AI, you have to bucket yourself into one of the four buckets or like five buckets or three buckets. It doesn't matter how buckets it is.

Speaker B: Yeah,

Speaker A: at the end of the day you'll have to say, okay, so the company is fully trained on you on AI and we're using AI agents, so we're kind of on a bucket three or the company. Yeah, some of the people are using AI and but it is not a company wide adopted like initiative. So we're kind of in the, in the nascent, you know, bucket or stuff like that. Um, and you know, you've, you've talked a lot about digital maturity and you know, you have your own skepticism about AI and you're not using AI a lot. Um, so we definitely, like me and you, we definitely fall into two different buckets when it comes to AI. Um, but for companies, because we, you know, we address companies a lot and we, we're talking with our clients, I think it would be very, like, very efficient and just good to have such definition to say for the company. Okay. So you know, your company would benefit if it is Indonesian, uh, or your company would benefit if it's in the connected, where you have, where you're building. 1, 2, 3, 4. Because as we said, digital maturity, it's not one more tool. I think in it, in AI, it will be just, it's not one more agent or not one more model. Like it will make the breakthrough. It's how you utilize AI through like a connected web of like stuff like, similar to digital transformation. And I will.

Speaker B: Not similar. Not similar. Not similar. Exactly. You nailed it. Like, yeah, that was. You couldn't have set up the next part of the conversation any better. Right? It's not different than digital maturity. Like a company, like whether or not a company is like benefiting from the use of AI. It comes down to many of the same considerations. Proper business alignment, leadership, you know, uh, executive sponsorship. Like, is it actually being rolled out within the company according to processes and with skills and expertise? Is it integrated with your data stack and the rest of your technology? Right. All those same considerations that we've hammered on related to kind of digital capabilities in general, all of that applies to AI. Ah, for sure, for sure. Um, I think in terms of just some quadrants that I'm thinking of, we've kind of talked about the whole generative AI stuff. We've talked about the large language models. Like most companies, that's not where they're investing their time unless that is your focus as a company. Um, but there's two other areas I do want to make sure we talk about. You just touched on them a little bit, the agentic AI. So these are you know, kind of semi autonomous kind of bots of one kind or another that you know, that are, you know, have directly act, ah, direct access to an underlying, um, model and they're able to autonomously kind of do certain tasks. Right. Like this is an area where um, companies that are a bit more at the forefront are definitely utilizing um, these agents. And um, the big difference is, you know, unlike a chatbot that has like you know, five rules or whatever, you know, if a person inquires about their order, you know, call up their order details. If a person inquires about their booking and says cancel, you know, ask do they want to cancel this reservation, whatever, right? Like those are kind of rules based, um, chat bots and they serve a purpose. But, but agents, they actually, um, because they have a more sophisticated underlying language model that they're taking advantage of, right. They're able to in a more kind of natural Language, conversational style, kind of learn what a user is looking for or learn what like based on what tasks they're assigned to do. They can get better and better and they can autonomously kind of like they're, they're from the standpoint of the person employing the agent, the AI agent. Right. The agent is focused on delivering the task, not being told how to do it. Right. So they, in that sense, um, it's a, it's a huge opportunity there in making like m. Potentially more sophisticated, um, tasks and deliverables increasingly automated. Like you're seeing it like on some of the, like even some things like the QA and development cycles on software, like some of that stuff can just be done by these autonomous agents. Right. Like, so there's some pretty sophisticated tasks that um, that can be um, handed off to, to agents. I think in the, you know, customer support, um, troubleshooting, um, you know, appoint even things like I know it's not very personalized but appointment setting and you know, if you call the hair salon and you have an agent just you know, oh, you want to come in on uh, April 25th, we have an appointment available at you know, 4:00pm like right. Like, but so that stuff's not that really that difficult and people are already doing it. I'm m. Not even. That's not even something that's possible. It is being done. Um, uh, I mean some of these things are doing like medical diagnosis too. Right. Like in terms of. And they're actually pretty like even like. Yeah, I mean some of this, some of the um, like say looking at mammography, uh, scan MRIs, brain scans and things. Some of the, some of those models are pretty darn good at calling out anomalies and potential um, health issues. Like in ways like humans are, are also susceptible to errors. Right. And misdiagnoses and stuff. So some of this stuff is super, super, super interesting. Um, and I do think in terms of like say use like nodding back to sort of Shopify CEO like they're in sort of the E commerce space, um, online travel. Um, there will be a, ah, there already is a push and there will be an increasing push and investment in uh, leveraging agent, you know, AI agents to do it more and more, no doubt about it. And also just like repeatable tasks. Um, there's definitely no doubt about that. Like it's just that that's happening like whether we like it or not. I have mixed feelings about it, but it's not like it's going backwards probably. Um, and then there's the other place we, we haven't talked about at all yet which is um, much like sort of GPTs, um, all of the native, you know kind of AI native apps and embedded AI functionality that's part of almost every like, it's increasingly part of like every technology platform. So most marketers, advertisers, content people, um, companies like if you're using Office365, right if you're using ah, Google uh, Suite, you know for um, um, your Gmail, corporate Gmail and things you're probably already using AI even if you don't know it, right? You get little writing assistance popping up when you compose an email, you get a little writing assistant, it asks you if you want it to like make it better. Uh, if, if you start searching for things in office365 you know uh, co pilot is going to try and kind of step in and help you, right? Like there's all kinds of things that um, if you advertise in Google Ads there's a whole bunch of AI ML behind the scenes and some of it's increasingly uh, also within the UI itself, right? Surfacing insights. You can do natural language queries in, in Google Ads you can do natural language queries in Google Analytics and get insights responses. Right? Um, most platforms are embedding these kinds of features, functionalities underlying algorithms, models like um, so even if you're not investing in AI as a company you're probably using tools advertising in places. If you use customer engagement platforms even though you're not like personally employing AI probably if you're using any of the automated journey and campaign optimization around um, automated triggers optimizing like do you send SMS vs email do you send uh, know time of day, you know automatically optimizing time of day like a lot of these or you know different versioning um, whatever those different levers are that get pulled by say a platform like Moengage, uh in optimizing your customer engagement channels. Right but like you're using AI, you're just not direct, you're just not directly and personally employing it. But, but the tools you're using or if you have like you know, some many kind of bigger, you know, kind of, you know, kind of mid tier and enterprise companies are Salesforce shops. And if you're using Salesforce and you're taking advantage of any of those things like you know, Einstein, right, Like these insights engines that's all being driven, you know behind the scenes by sophisticated models. So yeah it's, you're, you know I think you know to your point, Diode, right. You could kind of. It may not match exactly to the current digital maturity framework, but conceptually very much you can fall into these different quadrants. But even if you're very, very nascent, the likelihood that you're, that your people and your company are not utilizing AI and ML or ML and AI at this point, um, it's pretty unlikely unless you're um, unless you're using hand saws and um, and ads is to like to, you know, to carve like handcrafted wood furniture and you're not selling it online. Right. If you only sell it out of your sort of, uh, garage or uh, you know, local retailer and you don't use the Internet at all, then, you know, quite frankly, maybe you're not. But, but not very many businesses fall into that category at this point.

Speaker A: So. Yeah, I mean, I'm fairly interested to see who's gonna be the first one to come up with the framework of like implementing UI to like to work in your company. And like, how can you assess, uh, which stage you are in and which stage would be optimally best for you. That would be very interesting, to be honest. I mean.

Speaker B: Yeah, we should, we should come up with a framework. Uh, let's do it. I think, uh, I mean there's already some people I've seen. Scott Brinker has pretty, some pretty interesting stuff on it, um, over at Chief Martech and I've seen a little bit, but I don't think anybody's really just come out and planted a flag like this is, this is the, the AI, uh, maturity, you know, framework. Um, and I guess in each of those sectors there would be different, there would kind of be a different quadrant too. Like if you're in the companies that are actually developing the models versus, you know, ah, like, so we almost have to verticalize it. Right? Like, sort of like what's the AI maturity for retailers or for E commerce or for, you know, Omnichannel? Hey, that's not a bad thought. We need to do that.

Speaker A: Yeah, that would make up for a killer blog, to be honest.

Speaker B: Heck yeah, we should do it. We should definitely do that. And how about, could we get an AI agent to give us better personalities, make us cooler?

Speaker A: I mean, we're cool as is. We can use NotebookLM. They have some interesting, like, sound AI models that, you know, they're good at mimicking how people would interact and they're pretty fun, to be honest.

Speaker B: Yeah, there's a couple of, um, video generation, uh, platforms that I'm kind of Interested in, to really do the full featured aspects of it like text to video, like in, in like really powerful. They're a bit expensive the licenses at this point, so I've been a little reluctant to. But it's something from a research standpoint. I just want to, I just want to really see like how far the really good. The best ones have. Have come at this point. But I haven't been willing to, to just drop my, drop my credit card down and, and find out. But um, even some of the basic ones are pretty interesting.

Speaker A: But I mean they're, they're kind of not bad. I mean I think they, that's kind of.

Speaker B: Yeah.

Speaker A: Some time to cook.

Speaker B: Yeah but they're not. I mean to do a really, really, really good deep fake like other than just like a photo. I think on the photo end of things it's pretty. Gotten to be pretty darn easy. But like the, what do they call it? The uncanny valley. I mean like even some of these really, really expensive, some expensive ads, like people are just like not responding to them well. Right. Because they kind of, it's just sort of like there's something about it where you're just kind of like eh, like you know, like something about your visual response to it. You're just kind of like it's not real.

Speaker A: Yeah.

Speaker B: Yeah. So.

Speaker A: Yeah.

Speaker B: Which is a good thing I think

Speaker A: actually it's a good thing to get better. And um, I, I don't think we will be able to differentiate between what they are and what's not. And that's where we will, shall we shall live in the uncanny valley because we'll start to just like be skeptical of everything. You see.

Speaker B: I know. That's when we need to start rear. That's when we need to start rereading um, European critical theory, um, from the 20th century, you know, like uh, um, Jean Baudrillard and, and Herbert Marcuse. So like simulation and simulacra, like if you can't tell the, if you can't tell reality apart from the, the mimicry. Man, we're in a, we're in a crazy world at that point I saw

Speaker A: me and like a couple of other like uh, companies have been implementing like a watermark into AI generated images. And I think this just needs to be wildly adopted because you know, be able to just like differentiate between what's, what's AI and what's not.

Speaker B: That's good. So not everybody has sold their soul yet. So uh, it's good to see companies like the size of Sony committed to Identifying and flagging, ah, AI generated, you know, content to say this is, you know, um, it's important. I, I agree. It should be widely adopted. Um, yeah, they say there's, they say that recently some of the tests, you know, the touring tests for whatever, you know, can a human, uh, recognize that something was, was generated by a computer? I guess it's getting, uh, getting harder and harder for most people to distinguish the best, the best, you know, the best generated content. It's becoming like, it's like in this high 60s and low 70 range of people who think it's, it's real, legitimate, like human generated content. So yeah, it's scary times in, in that way for sure.

Speaker A: Definitely. All right, so closing thoughts, closing thoughts.

Speaker B: If you're not using generative AI in some way at this point, like um, as an organization, you're like woefully probably behind the curve in terms of sort of adoption, um, and like just take it for granted. You're. Even if as a company you don't have a policy or a, a process or a commitment to it, your people are using these tools. I'm, um, not saying anything very controversial there. Um, also, again, even if you think you are not investing in AI, if you are advertising in um, paid social, in paid search, um, if you're advertising on LinkedIn, if you're using customer engagement tools, if you're using most corporate productivity suites, you're also using um, embedded AI and in some cases applications that are just AI native. Um, so you're using AI in that regard as well. Um, and then, you know, the, the, you know, the question becomes, you know, like, are you, do you have certain kinds of customer experience, user experience, uh, internal processes, you know, maybe that these kind of more sophisticated, uh, AI agents would be an appropriate kind of an investment. You don't have to develop them yourself. Like, there are actual people developing these agents that can be trained or already are, ah, pre trained, whatever. Right. But um, so you don't have to actually build the agent yourself, but you might as a company benefit from that and then that other quadrant of actual, like the nerds actually developing this stuff, uh, like most companies, you don't need to necessarily. It's nice to know how a large language model works or you know, a neural network and all these things. But like, honestly you're unlikely for most companies and brands to be invest like actually investing in building those things.

Speaker A: So.

Speaker B: Yeah, you know. Yeah, yeah, yeah.

Speaker A: I mean, I would say, um, let them do the heavy work. They have the budget for it and just, like, read the benefits. Yep.

Speaker B: Cool. Well, we have entertained, uh, ourselves yet again.

Speaker A: We did.

Speaker B: I will say we haven't actually done it. We haven't had a old school, uh, episode where it's just you and I pontificating about digital things in a while. We had a lot of good.

Speaker A: Yeah.

Speaker B: Guests. We've had you, Sean and I, uh, doing episodes. But it's nice to just, you know, reconnect and have a. Have an old school, uh, soapbox.

Speaker A: Yeah. Back in my old days. That's right.

Speaker B: All right, brother.

Speaker A: Well, it's good to be back. Yeah. Yeah.

Speaker B: Enjoy your evening and we'll do it again soon.

Speaker A: Yeah, definitely. We'll see you later. Thank you, Patrick.

Speaker B: Thank you, dude. Take care.

Speaker A: Yeah. Bye.

Speaker B: Bye. And that's a wrap for this episode of the Digital Disruption Podcast. We hope you enjoyed our deep dive, uh, into the world of digital transformation and gain valuable insights to help navigate the ever changing landscape. If you have any questions, comments, or suggestions for future episodes, we'd love to hear from you. You can reach us on our website, esens.com or connect with us on social media. Don't forget to subscribe to our podcast and stay tuned for more exciting discussions on the forefront of digital innovation. Remember, the future is now, and it's time to embrace the power of transformation. Thanks for listening, and until next time,