EP 139: How To Transition to an AI-First World

This is the Everyday AI show, the everyday podcast where we simplify AI and bring its power to your fingertips. Listen daily for practical advice to boost your career, business and everyday life.

When will the time come when our work lives are just infused with AI and it's everywhere we go in our, in our work, in our business and our companies. Are we a decade away? Maybe a couple years, months? And then how do we transition to that AI first world? We're going to be talking about that today and a lot more on everyday AI. Welcome. My name is Jordan, I'm the host and I guess you're your guide to learning and leveraging AI. So I'm extremely excited to talk to our guest today about how we can transition to that AI first world. But before we do, as always, we're going to go over the AI news for the day. And hey, if you're joining us on the live stream, thank you. Let me know where you joining us from. I always like to see that and to engage with our live audience. If you're listening on the podcast, thank you. As well, make sure to check out your show notes. We always leave a link to go sign up for our free daily newsletter as as well as other relevant episodes on today's topic. So make sure you check that out. All right, so normally we run down a whole list of things in AI news but pretty big news yesterday. So we're just going to focus on actually one thing. So if you want to get the rest of the news, we got it, don't worry, just go to your everydayai.com and sign up for it. But yesterday big news. So let's recap it. We had the Open AI Developers Day and a ton of news when it comes to OpenAI and ChatGPT so for developers. But I'm going to break it down real quick in this little quick news segment on what it means for everyday people because it's not just for developers but high level. Here's what happened at the OpenAI developer conference. So they announced GPT4 Turbo, which is their newest and kind of most refreshed model which is supposed to be much stronger and a way longer context. A couple other things is being able to for developers to have better knowledge so being able to bring in outside documents by default in their products that they that they build as well as new modalities which we've seen these kind of already. So you know Dall E3 Vision, the ability to upload documents and to be able to work with those and chat with those. So you know now chat GPT developers or GPT developers have that ability to bring that kind of functionality to the products that they build. So a couple things that is going to affect just about any GPT related product that you use. So get ready for a lot of changes there. But a couple other, I'd say three, actually two, two big ones for everyone else. So we have copyright shield and that's going to be mainly for enterprise and API customers. But OpenAI is following suit for other companies like Adobe who are offering some of that copyright protection. So if someone gets sued for using their product, they will offer a sort of legal protection with the copyright shield. And then we have as well. It's, it's going to be cheaper, right? That's the other thing, about two to three times cheaper. So maybe those products that you use that are very expensive might come down in price. And then last but not least, chat GBT some big updates. So yes, the default mode for GPT4 will be Turbo. So, so faster, better and a much wider range of memory for ChatGPT. So hopefully fewer hallucinations as well as GPTs, which are custom GPTs that you can build without code being able to upload documents, train it to respond a certain way. So very exciting news out of the developer day and again, more on that and a lot more on what's happening in the rest of the AI news world. Go to your everydayai.com but you didn't come here to hear me rant about ChatGPT and Developer Days. You came to learn how to transition. You're listening because you want to know how to transition your company or even yourself to an AI first world. So I'm very excited to bring on our guest for today and please, if I can get it right here, help me. Welcome to the stage. There we go. We got her. Tina Yazdi is the founder and host of the AI First Business podcast. Tina, thank you for joining the show.

Hi Jordan, thanks so much for having me on.

Hey, I appreciate you. It's always fun to have another podcaster on the show. I've done it two or three times. It makes for fun conversation because we just chat about AI all day. Right. So tell me maybe a little bit, Tina, about the AI First Business podcast and what kind of your focus is there.

Yeah, for sure. So the AI First Business podcast came about as a conjunction of a few points in my background. One is that I have a philosophy degree specialized in AI. I also worked for a number of years selling a product that implemented into the data stack of startups and big tech companies. And that gave me a combination of exposure to some AI hygiene and also some real field experience on what the data stack actually looks like behind the scenes of these like companies that we put on pedestals. And in the last year those two things kind of collided together where I'm, I was witnessing a lot of narratives around the upside down inevitability of the impact of AI on our world. And I don't agree with that. I think there's a slightly different narrative that I think might be a bit more pragmatic and closer to reality, which is that the transition is going to be fast but slow and organizations are going to need a lot of time and a lot of unsexy admin work to be in a position to truly implement an AI first habit in their organizations.

Tina, you bring up a great point because I think people think of AI as something sexy. They're like, oh yes, look at this, it's the bell, it's the whistles. It's going to make everything great. But to get there as an organization, it takes unsexy work. You got to do the admin updating your, your governance, your, your legal. So talk a little bit about challenges that, that you see businesses are facing on getting to that point where you can actually start implementing it.

Yeah, absolutely. So my, my background is working with enterprise and large, like multinational groups in general. So this is kind of more the angle that I'm coming from regarding this. I think the first thing is just having the data available at all at this moment. I can pretty confidently say that most organizations don't have the quantity of data to run AI algorithms with any level of reliability. They're doing it more as a game of novelty, which if that's the introduction to get your team like comfortable with having it available in your organization, like there is some merit to that. But in terms of bringing an output that you can actually make decisions on or improve your cogs, I don't think that's quite what's gonna be the outcome of that. The danger of that is also it kind of will already start to erode trust in something that has shaky trust foundations among executives anyways. And I think the other thing is the true load of work, like kind of in combination with that is like the data quality and the data collection. That takes a lot of like data pipeline building, a lot of unifying your data warehouses. And this means hours and hours spent by data engineers doing work that isn't actually AI yet. And I think those are some of the things that I don't really hear talked about. And I think you can't really skip those steps and have, you know, any, like, business results coming from AI. So I, I think one of the dangers of the cycle is that one, it creates a situation that kind of will lead to an outcome where executives are like, aha, I told you we can't trust AI or I told you this was useless. And it kind of like, you know, creates a cycle of despair. And the other is that it leaves such a big opportunity on the table where you can have more productive teams. You can have people working on things that they love to do and do less of, like the grunty work that they don't like to do. But without the foundation, that transition won't happen. I don't think.

Yeah, I think we're all looking forward to less grunt work and, and more meaningful or more exciting work. And. Hey everyone joining live. Thank you. You know, Shannon said, good morning, AI Curious World. And, and I'd like to know also, like, get your questions in for Tina. What do you want to know about kind of this AI First World and, you know, becoming an AI first organization and, and maybe let's start there. I'm going to rewind and know how I started. The top of the show is talking about when, right. So when will our business lives. And I'm sure it depends on, you know, where you live, right. Whether you're in the US or Europe or elsewhere, you know, how your business maybe transitions to this. But from your vantage point and with your background, Tina, what are you seeing? Are, are you seeing. It's, it's very far away. Until AI is infused in every part of, of our business day to day. Or do you think maybe it's coming faster than some might think?

It is really scary to answer that question because I need to draw a line in the sand for myself as well. I think I'm going to be, I'm going to stick to the. I think it'll be fast, but slow. I think the conversation around AI is unavoidable and everyone's talking about it. And thank you, Mike. I, he just mentioned that you can't really skip those steps. I'm kind of drawing a little my experience rolling out other types of data updates to organizations. And given how those processes go, I'm going to make some assumptions that it's going to be similar with AI transitions because AI, an AI transition is ultimately a data. It's a subset of your overall data strategy and data like, you know, vision. What I'm Seeing happening is that there is interest in like level one AI tooling. And in fact level one AI tooling has been around for years. I work in SaaS sales and we have a number of AI style tools that have been forced upon us like starting 2016 already I can safely say. So that's nothing new actually. And these are things that analyze your recorded calls and give you suggestions on how to improve them. They scrape the data that you have internally on customers and help you make better like either categories or decisions about how to manage those things like that those are becoming a little bit more advanced and those are the things that are being rolled out right now. There's already some blockers there. We just talked about this before the show which is that the security and legal teams will probably be the number one stumbling block to that which is very much their job. Their job is to maintain a level of conservative, sober analysis on bringing new technologies into the organization. But on the other hand I don't see them being very well prepared to do something as simple as sign a MasterCloud agreement with an AI service because it just has variables in it that they've never had to deal with before. And there's no pre existing examples that they can cost, you know, copycat to to have confidence that they're making the right decisions. And so I feel like there's a little bit of a cat and mouse game there and I think that's going to prevent companies rolling out these technologies even at the level one stage to keep up with productivity. I think younger companies and tech startups, and particularly companies in the US versus Europe are better positioned to move fast and get that competitive edge. But I think that's for like larger organizations. It's for sure going to be a blocker. And I think the other one is kind of like I think companies might underestimate the resources required and the maintenance to get your data and have the right team in place to build like sustainable long term algorithms that provide business value for years to come. Like that will require a lot of foundational work that I don't know if it's in the fiscal budget, if it's in the headcount, if there's leadership, they can put together a team and like a three year plan. I don't know if there's an appetite to have a long term vision which you cannot avoid to rule these things out. It's just going to have to be a long term game. So I think those are some of the factors that will influence the speed to market of AI first habits in organizations today.

Yeah. And Tina, I love what you said and I completely agree the fast but slow kind of piece to this because generative AI has been on the scene for a long time, Right. But it's been kind of quote unquote mainstream for a year now, right. Since ChatGPT was released last November, this kind of opened up a lot of companies eyes to the power of generative AI. So what kind of pieces do you see have to fall into place for this fast but slow kind of transition to an AI first world to take place? Right? Because we talk like, okay, legal team, there's, there's data, there's governance. What are kind of some of those key pieces that if someone listening to this podcast is maybe a small business owner or they're a decision maker in a larger company and they're not there, like, what are those pieces that they have to get moving together to get to the place where they can be ready when they're going to need the power of generative AI.

So in terms of the fast and slow, I talked about the slow. Let me talk a little bit about the fast, which is bottom up, and I'll talk a little bit more about the slow, which is top down. What I think will drive the speed here is your employees are users in the real world and they are users of technologies. They are like obviously also the same people who are using chat gdp. I think one of the things that is forcing the hand for organizations to figure things out quicker and is also providing the firepower behind things like IBM's Watson and Salesforce's Einstein is that whether you like it or not, your employees are going to try to do things in the most productive way possible. And that means they're already using generative AI that's in the wild right now and that is causing serious enterprise gap security gaps and concerns. And I think that kind of forces you to like, smart, smart prepared enterprises will have to acknowledge that they're going to be using it. So you might as well create a guardrail for them to use it within the like, liability framework that you're comfortable with. Which means, like, you need to figure out a way to find agreements to like, either purchase this and bring it internally or create some kind of guidelines to operate by. So I think that's going to be where the fast comes from. Like it's happening whether you like it or not. It's up to you whether you adapt or not or just like put your head in the sand. And, you know, I don't think that has historically worked well for organizations in terms of what I, what I've seen the best in class organizations get into place to have a working movement towards AI. First one is you need a dedicated team for AI that has both a short term ROI scope and a long term ROI scope. An example of this could be like for example, gaming companies. And what I mean by that is, you know, transitioning your data and your algorithms to take advantage of AI has an element of like an element of ambiguity to it. So it is totally of course appropriate that you have some short term ROI that you, you know, put the team against that like you, you need to see in the next year or so. But unless they also have a bit of a scope to play around and build longer term Systems to have ROIs, like three years from now, maybe not immediately in the next year, it's, it might be a little bit hard to see the true potential of AI unleashed for where you're heading as a business.

I love that point. And there's, there's such a disparity there between. Yeah, because I feel like for, for you know, medium sized businesses especially planning for long term ROI is the norm, but you know, preparing for short term roi, if I'm being honest. Yeah, I've been a, you know, a digital dork now for 20 years. I've been working in different, you know, marketing and tech and comms roles professionally for 20 years and I haven't seen anything, not even the Internet that can provide such a short term roi. It's hard to measure. I feel like Tina, like, and I know this is a tough question, right, it's the trillion dollar question, but I mean, what are ways that companies can, can kind of even gauge that like short term ROI from generative AI because it's so fast and we don't have really a roadmap to follow necessarily. Are you still running in circles trying to figure out how to actually grow your business with AI? Maybe your company has been tinkering with large language models for a year or more, but can't really get traction to find ROI on gen AI. Hey, this is Jordan Wilson, host of this very podcast. Companies like Adobe, Microsoft and Nvidia have partnered with us because they trust our expertise in educating the masses around generative AI to get ahead. And some of the most innovative companies in the country hire us to help with their AI strategy and to train hundreds of their employees on how to use Gen AI. So whether you're looking for ChatGPT training for thousands or just need help building your front End AI strategy to you can partner with us too. Just like some of the biggest companies in the world do. Go to your everyday AI.com partner to get in contact with our team or you can just click on the partner section of our website. We'll help you stop running in those AI circles and help get your team ahead and build a straight path to ROI on gen.

Yeah, I would say like I think one of the classic use cases or business cases for AI is just the automation one where you can just take this goes back to the grunt work. There's something that has a lot of grunt work or a lot of manual lift from particularly your developers and technical team. Can you do something with AI that increases the accuracy of the results that you're making decisions on and reduces the hours spent? Then the ROI is pretty basic. It's the cost per hour for the employees in question and reducing that, that's a really simple calculation that you can do. But also I'm now thinking a little bit like in the cyber security space, like if there's for example a certain service that you run to audit something for your cybersecurity clients. Right. Improving the accuracy like 1% or 2% like what is the business value to your client? Like what kind of risk does that reduce? What is the cost of that risk? I think those, if you break it down into its components, those are some ways you can calculate and holds AI teams accountable to short term ROI.

And, and you know, some actually fantastic questions here. Hoping, hope we can get to a couple. So Michael here with a great question. Thanks for your question Michael. So he said how do you see this disrupting the digital divide? So saying with everyone having access to chat, GPT and pro versions for pretty cheap, do you think it will empower the low end or do you see it empowering the big guys to dominate and drive up costs in the dividend increases? That's a great question. Tina. What's, what's your thought on this? Is this going to, you know, going to this AI first business world, is that going to drive that divide or help close it?

I think a little bit of both. I think what's really funny about this, I have some real examples but I will not name any names. Is that AI tools like Chat, GDP and the pro versions as well can be lumped under the general category of productivity tools like Asana. Right. And in this sense, when you look at it in this way, which some organizations do, they become a commodity, they become a very low value thing that is actually really hard. Again, I have a sales background. So I'm looking at this from a sales perspective. It's actually really hard to command budget from it. And in fact the way that you can actually influence organizations to invest in tools like that at a leadership level is simply by pointing out the, and kind of exploiting a little bit the danger of using open source tools to the organization and the liability that that costs and the chance that your makes sense to the Internet. So that's actually one of the only ways at the moment you can like get budget for stuff like that in terms of empowering like I think that it improves like personal productivity and productivity per employee and I think that's really powerful the way the different organizations value that ranges quite a lot. So I think it depends a little bit on the organization. I think nimble, fast moving organizations, tech companies, series A to D types of companies, they really will understand the point here. I think maybe some more traditional larger organizations will need some time to see the value that that can bring and feel like they're falling behind competition by not making the right investments here.

It's such.

I'm curious by the way what you think about this question. I'm sure you have some observations also.

Yeah. Oh thank you. Yeah. Normally people don't turn the tables on me because I might go on a rant. I'd say I, I don't think necessarily that business leaders are looking at ChatGPT as a productivity tool. I think originally they were but I think as you see this push toward enterprise and then when you see, you know, which we haven't even talked about much on this show, when you see Microsoft Copilot, I think the smart forward thinking business leaders are looking at generative AI as less, less of a productivity tool and more as an essential and a new way to do business. But I, I like that's a very common, I think, you know, not, not argument. But yeah, people think okay, is this just something that's going to increase efficiency, you know, productivity or is this the new way that we do work in general? And I think that's probably Tina what we're going to find out over the next couple of months. You know, as Microsoft Copilot is starting to roll out here over the last week and I think it's going to take a couple of months. But I mean. Yeah, what do you even think about that? What do you kind of think about my take on that? We'll go back and forth here.

I think, I think both are true. I think this is not an or, but it's more of an. And I totally agree with you. And again, like coming back to the sales perspective, like, I think professionally my what I find really fun to do is figuring out like how do you take something like this and pull out like the core narrative that fits like the particular organization that you're speaking to. And even if both our points are right, you have to figure out but which one is going to resonate in this situation. And I think that's where, where things get really interesting.

Yeah, yeah, yeah.

And I think this is like moving in real time. Right. I don't think organizations like there's no answer or playbook on how people are thinking about this. I think it's being formed through, you know, listening to podcasts like this one, through everything you're reading every day, through conversations with your colleagues. And it is not in any way concrete. It's very much water still. So that's.

Yeah. As soon as you feel, you know, as a business leader, maybe an entrepreneur, small, small, you know, business owner is, I feel as soon as you feel you, you have it, like you haven't understood massive updates like with what happened Yesterday with, with OpenAI's Developer Day, that shakes up that, that long term plan at times as well. So maybe, maybe we'll focus on this something. A great question from Mike here. Mike, thanks for your question. Thanks for joining us. So Mike saying Tina, please share AI benefits in enterprise that are low hanging. Right? Because sometimes when you think of that long term and you see all these new tools and advancements and this and that, it's like, oh my gosh, how are we going to implement this? But maybe for those in enterprise that are maybe looking for that shorter roi, what are maybe some tools, tips or processes that can address more of that low hanging fruit that you can really start to measure for sure.

And again, like, I'm not saying this is the answer. This is just an example that comes to mind. And in a year from now I might look back and be like, that was a terrible answer. But what I have for you right now is that I think bringing in something as simple as a generative AI copilot into your organization today is pretty digestible. It's not super risky. And it also de risks the reality that your employees are using generative AI tools, whether or not they are honest about it in the wild anyways. And I think because it's kind of like a, like most people are like kind of able to get their head around generative AI and may have tried it themselves, it's digestible enough that it's a good sandbox to give that first pass of how are you going to get it through the door. So it's a good, like maybe target practice. If you can maybe look at it in this way. On how were your, how will your legal team look at this? How will your security team look at this? How will your employees accept this? Like, I think the other thing that I don't see acknowledged explicitly enough is that people don't trust AI at all. It causes mass anxiety. There is big feelings about it and I think it's important to acknowledge that and understand why. Is it because people are scared to kind of take their jobs? Are they. They just don't get it and they're suspicious of engaging with it because it's like the end is nigh. There's a lot of big feelings about it. I don't know what it is for each individual, but I think it's important to acknowledge how your employees feel about it because you might make this huge investment and see like the reason it's implemented. It got approval, you got budget, all of it looks good. It's been rolled out through the entire infrastructure of your organization and no one uses it. They completely just refuse to engage with it. Those are things you can't really predict per organization. And I think to summarize, bringing in a pretty simple generative AI copilot that's built on your internal systems and seeing what happens is a good, low hanging, safe, ish way to test what it might look like for your organization and where the gaps are for, for your group.

Yeah, I think there's going to be a lot of co piloting. Right. Whether, whether companies actually using Microsoft Copilot or bringing, you know, someone on in a role that can just help them from top to bottom is, is so important to transition to this, this AI first world. Right. Because if you just throw this responsibility onto someone that's probably already overwhelmed, it may not, it may not turn out very well. So Tina, we talked about a lot in today's show so far, so everything from approval and budgeting for generative AI to the unsexy work of implementing it, measuring the long term for short term roi. But maybe if there's one, as we wrap up today's show, maybe if there's one takeaway that you really want people to be able to take that and to be able to use it and to help them really transition into this new AI first business world, what is that kind of one major takeaway that you want people to have Yeah, I.

Think as an individual and in regards to your career, like be curious and just try stuff like get the free version time box 30 minutes and just try it. Because even if it's scary for you or you're suspicious of it, just having more of a playful mindset in a way that's maybe not at work, but something at home like chat GDP is worth the investment because these tools are not going to go anywhere anytime soon. And if you are a business leader, maybe I would say there needs to be a little bit more conversation between technical teams and leadership about what's needed from bottom up and top down to if bottom up. Your engineering team, for example, is upset with you that you're not understanding the need for this amazing algorithm that they built. Maybe having an honest conversation about what is the executive point of view and what is the trust gap that's blocking the final headcount approval that they need from you. And in terms of the other way, what do the executive team need to be aware of to right investments to have a long term ROI from AI? Like, you know, sitting down with your technical team and being like, you know, let's get real, like, how many months do you need to get our data cleaned up? Like how many, how many months of data collection are we missing to be able to run an algorithm with any value? I think those are some conversations I see only starting to take place very slowly, but I think they're the hygiene that might be missing to make any real transition.

Well, I think whatever the case, the transition is coming fast, I think. Right. Slow. Slow, but fast. I love kind of that we've been able to talk about that multiple times. But Tina, thank you. Thank you so much for joining the Everyday AI Show. Really appreciate your time and your insights. Thank you for coming on.

Thanks for having me, Jordan. It was great to connect with you.

All right. And hey, as a reminder, there's always more in the newsletter. We're going to break down a lot more on today's conversation. You can check a little bit more of Tina's work as well. So make sure you go to your everydayai.com Sign up for that free daily newsletter and we hope that you can join us back again tomorrow and every day for more Everyday AI. Thanks y'.

All. Thank you. Bye bye. And that's a wrap for today's edition of Everyday AI. Thanks for joining us. If you enjoyed this episode, please subscribe and leave us a rating. It helps keep us going for a little more AI magic. Visit your everydayai. Com and sign up to our daily newsletter so you don't get left behind. Go break some barriers and we'll see you next time.