Leveraging AI

134 | AI Planning For 2025: A Blueprint For Successful AI Implementation

• Isar Meitis • Season 1 • Episode 134

Share episode

2025 is right around the corner, and if you’re a business leader, you’re likely deep into planning. But here’s what’s new: for the first time, AI is no longer just an option—it’s a necessity. Whether your company started integrating AI in 2024 or hasn’t yet begun, one thing is clear: planning your AI strategy for 2025 will be a critical factor for future success.

This webinar is designed to guide you through every step of AI planning. We've aggregated insights and data from top sources and combined it with our experience with hundreds of business leaders, to give you a comprehensive blueprint that covers all aspects of your business:
- Assessments to perform before the new year: Know exactly where your business stands in its AI readiness.
- HR strategies: Should you hire, train, or consult to get the AI expertise you need?
- Tech stack decisions: Which AI tools and platforms are essential for your business growth?
- Financial planning: Budgeting and allocating resources for AI implementation.
- Identifying and prioritizing use cases: Ensure that your AI investments deliver the highest ROI.

We’ll also walk you through the immediate steps you can and should take before 2025, as well as the ongoing actions you should prioritize throughout the year.

AI is moving faster than ever, and 2025 will mark a turning point for businesses that adopt AI strategically.

About Leveraging AI

If you’ve enjoyed or benefited from some of the insights of this episode, leave us a five-star review on your favorite podcast platform, and let us know what you learned, found helpful, or liked most about this show!

GMT20241017-154220_Recording_avo_1280x720:

Hello and welcome to leveraging AI. The podcast that shares practical ethical ways to improve efficiency, grow your business and advance your career with AI. This is Isar Meitis your host and this is another live episode of the podcast. And this is a special episode. Because what we're doing today is we're doing a webinar that will help each and every one of you to plan your AI implementation for 2025. Now, I want to take you a little bit back on how I've put this thing together and why I've put this thing together. So how is pretty simple. I work with multiple companies through consulting. I work with a lot more companies through my training and education programs for AI. And so I regularly. And I regularly speak on stages where hundreds of business leaders attend. And then I get to talk to a lot of them through the social mingling that always happens before or after these events. So I get inputs on AI implementation across multiple channels from leaders in multiple organizations. Large and small. So companies as small as, 10 to 20 people to all the way to large corporations, fortune 2000 kind of companies. So to talk a little bit about the larger companies, I participated as one of the speakers in the AI realized conference about two weeks ago in San Francisco. That was all fortune 2000 companies. And he was A lot of inputs today comes from there, but I'm also speaking most of the conferences I speak at actually has mostly small and medium businesses. And so I get a lot of inputs from over there as well. In addition, I've done a lot of research for this specific episode where I basically aggregated information from every credible source that I could find. So these are. Surveys and summaries that were done by consulting companies like Boston consulting group and Accenture and McKinsey, and these kinds of guys, PWC, as well as small businesses, journals that have aggregated information about AI in the past few months. And I was trying to put this all together for you. So together we can understand. A what's the status of right now, where companies are right now. And B what are the things that I see for the future where I see companies being successful, where I see companies failing, and what are the steps that you can take right now in order to be more successful. And when I say more successful, a lot of the things we're going to talk about, yes, they're planning for 2025, but the reality is you can, and you probably should start right now. So don't wait for 2025, you still have a full quarter to go. And but it's from a planning perspective, I chose to do this right now because many companies are in planning season right now, both from a budget perspective, HR perspective and everything else. This is like how this thing all happened. I, as I mentioned, I work as a consultant with many of these businesses. And so we're working on this planning right now. And I thought of sharing a lot of what I'm doing with them with all of you. So with that, as an intro, let's get started. I'm going to share my screen for those of you just listening to this. As the podcast afterwards, first of all, if you really want to, you can always go and find us on our YouTube channel, the multiply AI YouTube channel. You can go and find us over there and then you can see the slides as well. But if you are like me and you like to consume podcasts while you're driving or walking your dog or running at the treadmill or something that does not allow you to look at a screen, that's perfectly fine. I'm going to share, everything that's on the screen. So I'm going to share my screen now and we're going to get started. as I mentioned, I'm going to start with a lot of statistics of what's happening right now. So data management issues. So we're going to talk about two Really big aspects of the biggest issues that companies have right now in implementation. One of them is data. And the other is the HR and human side of things. But this is, per statistics is 55 percent of companies have avoided implementing different AI solutions because of data related issues. I actually think, and you'll see, I will share my thoughts on a lot of the stats. So I took the stats as is, I didn't manipulate them because if McKinsey thinks that's the statistics based on interviewing, 2000 CEOs, I'm not gonna, not show you the stats that they're sharing, but I'm going to tell you what I think from talking to again, hundreds or even thousands of business leaders, but 55 percent avoid certain AI use cases due to data related issues. I think the number is significantly higher than that, but either way, it's more than half meaning more than half of AI projects of company want to implement. They're not implementing because of data issues. So that's one important data point. 77 percent of organization acknowledged. They must implement new information management measures in order to benefit from AI. Now I want to take that a step back and explain why this is such a critical point. It is. Such a critical point because Companies have invested for years, whether you're a small business or a large business in building IT infrastructure and data management platforms. This could be in small businesses as simple as having your Google drive or your, SharePoint built properly. So everybody knows what it is and everybody has the right permissions and so on. And in large organizations, obviously a lot more complicated, but in addition, you have data in a lot of other external sources. Like Salesforce or HubSpot or your ERP system and so on. And having the data, the way it's built today is not helpful. And it's actually sometimes the other way around. It's preventing you from benefiting from AI benefits. And we're going to talk in the second half. Of this discussion about what you need to do about that. But again, if you are experiencing this kind of issue, know that three out of every four companies at least Are having that situation and I can tell you that in small and medium businesses, which is most of my clients It's a lot worse data life cycle management It's another big issue. And again, 75 percent of company have already increased their budget in 2024 for better data management. So again, if you haven't so far, this is something you start need to thinking about for 2024. And we're going to get to all of that. There's going to be a segment of planning and exactly what you need to do, but I'm just sharing with you, what is the situation right now? Risk management readiness, 77 percent do not feel that they're prepared from a risk management perspective to implement AI, which means again, three out of every four business leaders feel that way from my personal experience of talking to people, the number is much higher than that, and the people who are not thinking this way, who think that they are doing Okay. Usually just didn't ask the right questions. And when I, once I start having conversations with them that are a little deeper, they're like, Oh my God, what about this? What about that? And then they call other people. And then we started getting deeper into that. But I think the number, the realistic number of people who do not fully understand the risk management aspect of AI implementation is probably. 85 or higher. What are the things people feel ready for this particular statistics that I'm sharing with you right now is from. Large organizations, so the larger corporations in the world, 45 percent feel that they are ready from a technology perspective, 41 percent from a data management perspective, 37 percent from strategy, 23 percent from risk and governance, and only 20 percent from a talent and HR perspective, which tells you that one of the biggest gaps there are, there is right now is employee engagement. Training, education, and talent acquisition with AI knowledge across. All the different aspects of the organization. So when people talk about HR gap, and we're going to talk a lot about this, it's not just, Oh, I don't have somebody who can train models. We're also talking about every person in the company, all the way from your board to the CEO. Yeah. People who just you hired two weeks ago were in the trenches in the lowest levels of the company. Every one of these people needs to be trained about do's and don'ts of AI on what's possible and what's not possible on guardrails, on what tools you can use, on what are the use cases and so on and so forth. And that's currently probably the biggest gap that companies have across the board. And that's true, large and small. This is a very interesting piece of statistics, 67%. And this is again, a large business. more on the corporate side, 67 percent are increasing their AI investment next year due to strong early value from tests that they've done this year. So first of all, this is very promising, right? This is early, already early signs that companies that started early and are investing significant amount of money are going to invest more. money moving forward because they're seeing good signs of value from what they've already done. That being said, I'm not a hundred percent sure that information is accurate. And the reason is, and this is actually the next two pieces of information come from the same survey as this one. And now we understand why I question this early value thing. Only 41, sorry, only 41 percent of business leaders say that they're finding it hard to measure the ROI of AI. So 41 percent are saying they don't know how to measure ROI. And I call bullshit on that. And the reason I call bullshit on that is because the number is significantly higher. I talk to business leaders across multiple companies, as I say, weekly, whether speaking on stages and then talking to people, I've been in two large conferences in the last. We can a half. So I've met with hundreds of people just in that timeframe. And I teach my courses in which there's dozens of people, every single course, the number is significantly higher than that. But to prove that the number is significantly higher than that in the same survey, and I think that's PWC, only 16 percent of companies regularly produce AI ROI reports. So if only 16 percent produce ROI reports, to It's only how do 40 percent feel uncomfortable with their ROI and 60 percent saying they know what's going on. There's no freaking way. So the problem with ROI measuring ROI when it comes to AI specifically is that AI impacts every aspect of the business. So you have people in sales, in marketing, in HR, in operations, in customer service, in customer support, in planning, in strategy, data analysis. All of those are using AI or can potentially use AI for various aspects. And in each and every one of those departments and in each and every one of those use cases, it impacts different parameters. And that's why it's very hard to figure out how to measure. The trick is to always start with a clear baseline with specific clear use cases, and then you have a benchmark for comparison, and then you can make the get to the conclusions that you're trying to get to whether this is actually working or not. And then, based on that, based on those results. Solutions, and based on those outcomes, you can decide how to move forward. What are the reasons most, first of all, most AI experiments fail. So 67%, two thirds of companies move only 30% of their experiments to production. So most companies fail in 70% of their AI implementation attempts. That's crazy high. I don't see numbers that are even close to this. And again, based on my research and based on the companies that I work for, I can tell you what are the main reasons these initial implementations and experiments fail. Number one, no clear definition of use cases. So many companies start with the technology in mind and they go ahead and they buy everybody, co pilot licenses or some other licenses. And they say, okay, go use this. And then nobody knows exactly how to do it because they didn't get any training and there weren't any procedures defined and there weren't any guardrails defined, but everybody's telling them to be careful with the data. So they don't do anything. And so no clear use cases, no clear procedures are the two main reasons why these things fail. Number three is training like their use cases have been defined and the procedures have been defined, but there haven't been done enough or specific enough training for The employees on how to actually implement. The things, and remember, this is a change management process, right? You're going to somebody who has been doing the same job for a while. They have their way, they have their system, they have their process. And are you asking them to do this differently? Change management has always been about getting people buy in and excited about the change. And if you fail to do that, the fact you provided people with the technology, you provided people with the safety guards, you provided people with the procedures, you But you haven't created the excitement and the training, it's not going to happen. The next thing is data and data security issues. So scaling from a test to a full implementation company wide or department wide comes with a lot of baggage of having the right data and the right setup and the right security measures, and that's sometimes not easy to scale. And there's other scalability issues beyond that, that are also limiting. So these are the reasons why so many. Things fail, but the very first one is people and their training and clear definition of use cases. Now let's dive a little deeper into small and medium businesses. and I see a lot of people here are in that category of, 150, 300. I'm now reading the notes here, 200 employees. So this is in the category of small and medium businesses. 97 percent of SMBs are planning for AI adoption in the next 12 months, where about 65 percent are definitely. working to implement AI and another 30 something percent are considering implementing AI. A hundred percent of who I'm talking to is in one of these two buckets. and they are definitely talking about implementing AI in their business starting now. That's an interesting statistics. The question was, have, and that's again, a small mini business question. Have you used AI in your business? 96 percent said that they were using AI in their business. And I call bullshit on that as well, because I talk to many business leaders and maybe they answer the question in, yes, I know Gina once answered an email using chat GPT. And now that counts as yes, we've used. AI in our business, most small to medium businesses still do not have clear implementation of AI. So they may have individuals who are using AI in the business for specific use cases, but more than 50%, and probably a lot more than that, don't have a structured formalized way on how to implement AI yet from a different survey, a much more logical statistics shows that, 36%. are using AI regularly, meaning they have specific use cases that they're already doing. The other interesting part of this statistics, which I find more interesting is that 51 percent of growing companies. So small and medium businesses who see growth are using AI. So in companies that are seeing progress and moving forward, more of them are using AI than the average in the general population. Now, there's a question here. When you advise, and that's a question from, the chat in zoom is our, when you advise on how to use AI in a business, do you recommend certain software? Do you tailor it yourself? Also, do you work with developers to do this? So we're going to get to that afterwards. But to answer that very quickly, it depends on the company and the use cases. I usually recommend, and you'll see that after afterwards in the recommendations to start with off the shelf tools, just because it's significantly lower investment. You can prove the ROI very quickly, and then you can decide whether you want to go into custom development for your company, depending on what additional value you think it's going to provide and the, Potential resources that you have in your company, whether in money or in HR and skills to actually do that implementation, planning to use AI. 73 percent of small businesses are planning to use AI in the next 12 months. So again, a very high percentage. From my personal conversations, that's a hundred percent. I don't know who the other 27 percent are, but again, I'm not going to change the stats that I collected from highly credible sources, 55 percent of small and medium businesses emphasize that training or lack of training of expertise are their main barrier. To AI implementation again, if you figure out that part, if you figure out how to get your employees, how to get your board, how to get your leadership, how to get people in your company to learn how to use AI properly, you are solving potentially the biggest problem. in AI implementations that companies are experiencing right now. So if that's what you feel, I don't know if you need to feel bad or not, but know that you're in the same boat with a lot of other companies, training and expertise. 55 percent of small and medium businesses, emphasize, as I mentioned, expertise barriers, and then I mentioned, I call bullshit on that because the, there are a lot more. Companies that are struggling with training their people, then 55%. And again, I would say that's probably in the 90 percent number from all the companies that I work with. Now I might be tinted because the people who approach me because they want to. To learn, meaning by definition, they don't learn. So maybe my statistics is a little off, but I think the number of companies who are struggling because they don't have the right training and the right expertise in the house is more than 55%. 58 percent of companies of small and medium businesses are going to third party vendors to get assistant. assistance in initiating the AI implementation and in getting training for their people. So again, this is more than, it's almost 60%. Three out of two out of every three companies ish are doing that right now. It doesn't say why the other ones are not doing it. Maybe they're not implementing AI yet, but most companies, definitely smaller companies are doing that right now. Let's look a little bit about challenges for adoptions and how small and medium businesses are seeing that right now. So the four main categories on the technology side is. not having sufficient infrastructure. So that's one. Number two is, data and privacy and security issues. Number three is lack of in house skillset. And number four is. integration with existing systems, right? So these are the four biggest issues that small and medium businesses are seeing right now as far as implementing AI, but they're also seeing issues when it comes to non technical issues. Number one, as we mentioned, and by the way, that's number one, overall lack of skills and insufficient training. That's the biggest problem right now. Number two, lack of understanding of the scope of AI. And the objectives. So that's still, I would say a training and education issue just on the higher levels, more on the strategic side of things. How are we going to approach this as a company? What are the goals we're trying to achieve? So what goals we can achieve? Can we do with the current resources that we have, but that's still an education thing, just on the higher level of people. Number three is employee. pushback. So that's still education, right? Employees are afraid of AI. They're afraid to lose their jobs. They, if they are on the data security side, they're also going to push back. And so all these things are all an outcome of lack of education and training. And then, these are the top three things on the non technical aspects of AI. There's a question of where I obtained that information that I'm sharing here, all the charts and the specific stats comes from me aggregating data from research that was done by. Lots of credible sources, which are mostly the big, consulting groups, Accenture and BCG and PWC and McKinsey and from academia like MIT and, Harvard and the big providers like Microsoft and Salesforce and companies like that. So this is statistics from all of them combined. Reasons why small and medium businesses haven't adopted AI. The two main things. The number one is they don't know how to define the right applications and use cases, which again goes back to lacking the right expertise. And number two is lacking required expertise. So these are the two top things that small and medium businesses around the world are saying what's preventing them in implementing AI. So we go back to the same thing again, and again, Top priorities that came from serving 2, 700 companies around the world, small and medium businesses, number one, upscaling, basically taking your existing employees and training them to know more on how to use AI. Number two, AI risk specialists. So how do we find people or train people to understand the risks of AI and implement this so it doesn't come to bite us in the ass basically, once we implement AI, if we don't have these things in place, number three, Periodic training. So out of 2, 700 small medium businesses around the world, two other top three are training and education related. Then we have data privacy, data governance, cybersecurity, model testing, and model management. So these are all coming below the training side of things. So that kind of shows you how critical. All of this is when it comes to training and so on. So I'm going to open parentheses for a second. I will take this opportunity to tell you something that is not happening regularly, but it's happening now. So I have been teaching the AI business transformation course since April of 2023, at least once a month, this year, twice a month, every single month. So hundreds of business leaders and business people have been through that process. Now, most of the courses that I teach are private courses, meaning companies and organizations invite me to teach their people. When I say organization, some of them are just nonprofits, but some of them are consortia of multiple organizations or higher level. Industry organizations, which underneath them, there are multiple companies. And then they invite me and I teach all of their companies and so on, but I've been doing mostly private courses because of that, I don't have a lot of time to teach publicly open courses. So we've been teaching the open courses roughly once a quarter. It's not roughly, it actually ends up being once a quarter. So the last course I've been in July, the one before that was April and the one before that was January. So once a quarter, there's an open course. The next open course that is open to the public that anybody can sign up to starts on October 28th. So a week and a half out from today, it's a four sessions of two hours each that are going to take anybody who joins the course from basic understanding of AI, all the way to how do I start thinking about strategy from a company wide perspective. But in the middle of all of that, there's a lot of hands on tools and use cases that you're going to learn how to do. So this is the course Gets updated every two weeks. So literally every time we run the course, there's new updates based on stuff that's happening in the AI. That's moving very fast. So if you are in that bucket of companies that needs this kind of training and you are looking to join a course that can get you Accelerating this process by a lot, then this is a great opportunity. And also because we're doing this planning session and I understand the importance of this time of year for the planning for next year, we're giving a hundred hours off to anybody who's going to join and participates in this session. A hundred hours off with promo code AI 2025. If you are watching this right now, you can scan the QR code to go and sign up for the course. If you're listening to this on the podcast afterwards, you can just open the show notes and there's going to be a link to the course. And you can sign up still the same promo code to get a hundred hours off the promo code is AI 2025. By the way, for those of you who are in leadership position in companies, as I mentioned, I do courses and workshops for businesses that are tailored. For the needs of the business. The smallest companies I do this with are probably 20 employees ish. The largest company I have right now is 45, 000 employees. and in multiple industries. So if you're looking for ways to get training that are custom for your company, that could be two hours, that could be three days, that could be online, that could be face to face. This is something that I do regularly. This is most of what I do. And I take my background as being a CEO for the last 15 years and pour my AI knowledge and capabilities into this to create this very specific training that is focusing on business growth, but how to leverage AI to do that. So I'm closing parentheses, a lot of that, and we've got to go back to our statistics, but just because we talked about so much training, I thought it's going to be important to mention this. there's a note on LinkedIn that I agree with a hundred percent that the biggest mistake is not starting. Like many businesses are afraid or think it's going to go away and it's just a trend. Like other stuff was, it's not going away. It's not just a trend. It is a, maybe the most transformative technology in the history of mankind. And. Not getting started is only going to leave you further behind. That being said, in order to get started, you need knowledge. In order to get knowledge, you need some kind of education. You can do it yourself, or you can sign up for different kinds of courses. By the way, not just my courses. There's multiple other organizations who offer, these kinds of courses. And it's just a great way to get started. So let's look a little bit about what are the benefits. So why would you even do this? Why do you want to implement AI? So this is a summary from multiple organizations. Again, this is Probably either McKinsey or one of those, and 34 percent do this to get, have seen improved efficiency and productivity. 12 percent encouraged innovation, 10 percent improve existing product and services, 9 percent improved costs, reduced costs, which to me comes with efficiency and productivity. I think it's the same thing, but I want to talk a little bit about number two and number three. So improved efficiency and productivity are obvious, but encourage innovation and improved productivity. Existing products and services is huge. And why does this happen? It happens because by gaining those day to day smaller efficiencies. You free people's time to think and to do more meaningful tasks like invest in innovation and what needs to be the next thing. And the same thing comes to improved products and services. So it's not only that internal things gets better. You can now. Drive the direction of your company, provide better product, better services, more innovation stuff to your clients. Meaning you're going to go your company that way just by implementing these tools. Even if all you did is save some people some hours a week and freeing them to do the more important, more valuable tasks and do everything faster. The next thing. Is behaviors that company sees as providing most value to, behaviors driving the most value for generative AI initiatives. Number one is deeply embedding gen AI into functions and processes. So this basically means that you need to figure out how to tie AI to the day to day of your employees. This is the number one thing that companies are saying. Number two is effectively management risks. Number three is deploying the latest technology. And number four is developing creative and differentiated applications using AI. So basically leveraging AI to build new stuff for your So you can serve them in better ways using AI solutions or AI infused solutions together with your existing solutions. By the way, all of these again, go back to training and education, like to deploy and embed AI into your existing processes. You need to know how to analyze existing processes, knows which ones can be automated with AI, know which AI tools to pick for that. All of that comes with proper education and training. So what are the main. Aspects of the business that company are already implementing, implement AI. The first one is obvious marketing. 34 percent of companies are implementing AI for marketing and sales activities. 23 percent are using it for product or service development. So again, what we just touched upon, provide better value to your clients while or through leveraging AI and combining it with the stuff that you're doing anyway, and then. The numbers keep going down as far as the percentage, but then it's I. T. and other corporate functions, services and operations still 16 percent software engineering, 13%. I think that's going to grow significantly. I think the reason that's a relatively low number. The software development side is just most companies are not software companies. And In software companies, I'm sure that number is probably above 90 percent and I started writing code. I don't know how to write code. I've never written code in my life, but I now build these small little applications that I can build with AI to help me in small things in the day to day. And that will become more and more common across multiple companies as we move forward. And these tools become easier to use. So this is what's happening so far. And these are a lot of key, important things. But now let's talk about what's coming in the next two to three years. First of all, AI agents is coming. So what is AI agents? AI agents are these tools that, and what you're seeing on the screen by those, who, those of you are here and are watching the screen is an agent. That's going to order books from you on Amazon, just by me asking for books that I'm looking for. But AI agents are very different than what we know today, as far as. chatbots and the difference are several. One is they can quote unquote think for themselves. They can take a complex task, understand what it means in the context of our lives or our business, break it into smaller tasks, and then go and execute these tasks one by one to complete the bigger task while evaluating each step and correcting and so on. So I think having a project manager that manages the project, but it's a AI agent and not a human that's doing it. Number two is that they have access to tools. So again, those of you are watching the screen right now, while I'm talking, you can see that it's looking for the books. It's searching for Amazon. It's searching inside of Amazon. It's finding the cheapest option for that particular book. It adds it to, the shopping cart, and then it's going to go and do the checkout and all of that, but the same thing, it can have access to every tool we have access to digitally. So anything you can access with a keyboard and mouse and internet connection, these tools will be able to use as well. And that's coming, and it's coming right now. and the only two gaps right now in agent implementation, and that's why we're not seeing it widespread yet. One is they're not consistent enough yet, which is obviously a huge problem. And two is Our trust, like how much do you willing to trust an AI agent run wild with your access to your tools and so on? Probably not so much yet. I assume by the end of 2025, we'll start seeing some very significant implementation of AI agents, but the ability is being developed by all the big players in the market right now. Another important thing that's happening is Lack or significant reduction in organic traffic. And that's going to happen for multiple reasons. The biggest reason is going to be the fact that more and more people are using AI for search. And even in Google, the top results now are AI generated, results, which just gives you the answer. So there's no reason to click on any links right now. Statistics from Ahrefs, one of the most commonly used SEO tools in the world, almost 97 percent of websites in the world get zero traffic from Google. So that's. Problem number two is that 64 of Google users right now, that's before there's more and more AI stuff, never click from Google to go to another website. They just stay on the Google search page and that's going to get worse and worse. And what that means for you, it means that you need, if you, if your business depends on organic traffic, you need to start looking for other traffic sources. very quickly, or you're going to start seeing bigger and bigger decrease in the traffic that comes to your website. And because you're depending on it on your results as well. Who knows? What AGI is, let me know in the chat, and who knows who this guy is, let me know in the chat. Those of you who are just listening, just think about AGI if you don't know AGI. But what I have on the screen right now is an image of Sam Altman, the CEO of OpenAI, one of the most influential people in the AI world today. Some would say the most influential person. And he was asked about, so AGI is Artificial General Intelligence. And he was asked, By two very successful marketers. What does AGI mean for marketing? And he said the following, and I'm now quoting, it will mean that 95 percent of what marketers use agencies, strategists, and creative professionals for today will easily, nearly, instantly. For will easily nearly instantly and at almost no cost be handled by the AI and the AI will likely be able to test the creative against real or synthetic customer focus groups for predicting results and optimizing again, all free. Instant and nearly perfect images, videos, campaign ideas, no problem. That's when you can hear a needle drop after you hear that sentence. But it's actually a lot more extreme than that. And the reason it's more extreme is because he was asked about marketing. But if he would have been asked about anything else, he would have answered about anything else. Which basically, if we reframe this, what he's saying is that 95 percent of whatever it is you're asking about will easily, nearly instantly, and at almost no cost be handled by the AI. All free, instant, and nearly perfect. That's the world we're going into. And they're anticipating this thing to happen in about three years. The reality is it's not a one and zero game. It's not that in three years, we're going to have it every new version that they release, which is every few weeks. If you consider all these companies together, there's new capabilities that get us closer and closer. And some of these are already there for some business functions. We can already do 90%, 95 percent of what the task was with AI, without human participation or almost free, almost instantly, that's something that as business leaders, we have no way, how to grasp. Okay. Or as a society for that matter, but that's where it's going. And that's part of the things you've got to start thinking on how to plan to, we're going to touch more on that shortly. The next guy on the screen is Dario Amadei, the CEO of Anthropic, who used to be a part of OpenAI. He was leading the team who developed GPT 2 and GPT 3. And then he left and started his own company. so those of you use Claude, he's the guy behind Claude. And he just released a Manifesto or call it whatever you want to call it on his personal blog and he's talking about what he calls powerful AI. He doesn't like to call it AGI because AGI has different definitions about different people. He basically said, I'll call it powerful AI. And what is powerful AI? And now I'm quoting from him. It's going to be smarter than Nobel prize winners across most relevant fields. That by itself, mind blowing. Number two, has all the interfaces available to humans working virtually. I told you before, access to the internet, keyboard, mouse, cameras, anything we have that we use daily, the AI will have access to. And it will know how to use and how to understand. Number three, it can be given tasks that take hours, days, and weeks to complete and then goes and does those tasks autonomously. So this is insane, right? We now used to stuff that replies to us in a second, but this is saying, take a project that is six months project, give it to the AI and it will execute it over six months, obviously giving you visibility to what it's doing. So you can intervene, but it can run the project on its own. It will be able to run millions of instances that can either run independently, each one doing different things or collaborate between them to create projects. Basically, companies or societies of these AI or these powerful AI capabilities to complete a huge amount of tasks, and you will observe information and will generate actions at roughly 10 to 100 X what humans can do. So all of this sounds science fiction and scary as freaking hell. So then the question was, okay, so when is this coming? And I'm really wondering to see those of you haven't read this, and I'll give it a few seconds. If you're driving and you're listening to this, just think in your head. When do you think this is coming? Those of you are in the chat, please write in the chat. When do you think Dario Amadei, the CEO of Anthropic think this is coming? So some people are saying 24 months, 2026 and so on. So he's claiming that this is coming in 2026. So within two years, now this is one of the most advanced knowledgeable people on AI on the planet. Thinking we're going to have this in the next years, this is insane and nobody's ready for this. But as business leaders and as business people, we have to start thinking about it and what does this mean to our business? The last thing before we go into the planning, which is why we're all here, but this is, was all very important background for that is humanoid robots. Those of you haven't watched the Tesla event, which may or may not have been rigged, but it doesn't matter. Humanoid robots are coming, maybe not in 2026, maybe in 2027, but they're going to be everywhere. They're going to be in storefronts, putting product on the shelf. They're going to be in assembly lines. They're going to be washing dishes and walking the dogs for us. And they're going to be working our warehouses and so on, putting stuff in packages and shipping it. All of these things are happening in the next two to three years. Years at scale. So how do we plan for all of that? So the first thing we already touched training and education, right? And training and education, I suggest doing in two steps. Step number one is. External support, get somebody from the outside, somebody like me, but it doesn't have to be me again, there's multiple other people do amazing AI courses who will do initial training and will also do custom training for different aspects of your business. And when I say different aspects, first of all, different levels. So your board, your executives, your directors and your employees, but also different functions. So salespeople did need different training than marketing people who need different training than data analysts. All of those need different AI training. So that's external support in the beginning and then external and internal. So a combination of those continuous education from now till forever, because this thing keeps on changing every few hours, there's a new capability and you need to build a mechanism within your company to be able to continuously educate everybody in the company about this thing. And we're going to talk later on about how to do that. But To give you a hint, that's going to come from an AI committee. that's one of the roles of the committee, make data an accelerator and not a barrier. The first thing you want to start doing right now is consolidate your data. What do I mean by that is I talk to many, mostly on smaller businesses. I must admit that have some of their data in Microsoft, some of their data in Google, some of them data in Dropbox, some of their data in Salesforce, some of it, it's part It's going to be very hard for you, especially in small businesses that don't have the capacity to do huge projects, to enjoy the full benefits of data in the AI age and both Microsoft and Google will figure this out. Meaning you will be able to gain huge benefits that are still very far from that on co pilot or Gemini, as far as integrating everything within their ecosystem, but in the next 12 to 18 months, they will meaning you'll be able to know. With simple prompts and most likely with voice communication, everything that's happening across everything, Microsoft or everything, Google, and so put all your eggs in that one basket. Any of those will do and. It will make your life a lot easier moving forward and you can start to transition right now. next one is best practices for data. Start figuring out once you decide on the platform you're going to host it on how you structure your data in the best way possible. Now start with moving forward. So naming conventions, how do you name documents? How do you name folders? What is the folder structure for each thing that you're doing? Where do you store your data? Is it stored? Because right now I can guarantee you many of your companies, each person in the company saves data differently. Some save it locally on their computer, some save it on a Google drive, some save it on Dropbox, some have a jump drive. They're like, I don't know. You have to find the one place that everybody saves to and have the same procedures, same naming convention, same folder structure for everything moving forward. At least your future, data will be correct. And then if you have the time, the bandwidth, you Or an intern that you can hire to go backwards and now align your historical data to that. Start doing that because it will give you a huge benefit. Once you'll go to the next step of implementation. Minimize channels. Right now I get to companies like, okay, how do you communicate internally? And the answer is everything. They have Slack. They have teams. They have phone calls. They have text messages. They have WhatsApp. They shot across the window. It's impossible to aggregate all the data into something that AI can make logic of. So try to minimize your channels to as least number of channels as possible. preferably digital communication. So if you can do emails and Slack, awesome emails and, Teams, awesome, but don't have 20 other channels on top of that. Record all your phone calls with your clients, internal, with your suppliers and so on. Why? Because AI is amazing today in finding gold as far as data and understanding what works better and what doesn't work better in unstructured data like call transcriptions. And if you don't record the calls, you don't have that. And you cannot make those calls. Benefits out of the data that you can have without adding anything significant to your company, data cleaning and consolidation. That's a bigger project for bigger companies, but if you really want to get in the benefits of all the data across all the different sources that you have transitioning from historical, traditional way of storing data to a data lake or a data lake house, and I'm not going to dive right now into what they are, is the way of the future. And the problem is it's a bigger project that requires Significant, deep knowledge that costs a lot of money, definitely worthwhile for bigger companies, probably not relevant for smaller businesses. there's a question about, G, GDPR in Europe and the same thing probably here with companies with HIPAA and so on. Every different company with every different industry in every different region of the world has its own limitations. I cannot give in this session advice to all of them, but you have to take these things into consideration. Of course. Many of the big platforms now cover for that. So many of the air platforms and the infrastructures both on, AWS and Google and Microsoft Azure have these things in mind when it comes to HIPAA in the US for healthcare or GDPR in Europe and so on. But you have to have somebody who's an expert on that do the analysis for you. Now A. I can help you in all these things in understanding what naming conventions you should use. Show it what you have right now. It can help you define the naming conventions and so on. So you can use A. I to help you in your data set up data security. It's a big deal. And even if you don't understand it yet, it's huge. And so you have to find a person, whether I hire him part time or hire a full time employee that is an expert on AI risk, because it's very different than traditional software and data management. And it might be critical for the future of your business. AI and get all the benefits, Training and education. We already touched about that. You have to train your employees on what not to do. With AI right now, about 80 percent of employees are saying they're using, they're bringing their own AI to work without reporting it to anybody, meaning you already have huge leakage when it comes to your data. And unless you train your employees on what the risks are, you're going to get hurt or you might get hurt. Clear AI guardrails have to be a defined and be known by every single person in the company. You have to reevaluate your IT security protocols and systems that are in place because the AI age will require new stuff. And again, for that, you need either external expertise, or somebody full time to do that. You want to set up. A whole new infrastructure to support that. And that means potentially reorganizing the way your data is set. So you can a benefit as much as you can from the benefits while reducing the risk. So it might be that your existing data structure needs to change. There's always consideration between. Open source and closed source. So open source, you can run within your environment. Some of the closed source tools now you can run if you're running them through AWS on Bedrock or through Azure, you can run it within your environment without having data leaking out. Again, you need experts to help you implement that, but it's doable. And then the last thing I would say about data security is make love, not war. In many organizations right now, you have the people in the company who wants to move fast. It's usually the marketing side and maybe the sale side. And then you have the CISO or the, CIO who are like, no. You can't use anything. You got to find the middle ground and make these people work together in order to make the best out of it. HR. First thing you need to do is a skills gap analysis. What knowledge do your people need to what skills they need to have for your company to be successful in the AI era. And where are they right now? And you might have different people in different departments, a lot, but you got to identify who these people are and where they are right now. So you can identify the training, which is step two. What training do you need to provide? How do you provide it as quickly and as effectively as possible? Number three, hiring with AI. I now work with several of my clients. We've developed tools to write job applications that are better, more customized to the specific needs that we have in specific markets that can, review open positions and optimize for them, that can review resumes that people send in for specific keywords and specific things that can write interview questions for the people we're interviewing that take into considerations the needs of the role, the resume of the person, the company culture and core values, the personality assessments that people submit. It can take all of these into account and write interview questions that will help you better identify whether it's a good candidate or not. These are huge. The flip side of that, most of your candidates will send you resumes and job applications that are written by AI. So they're going to be much better than everything you used to so far, and you won't be able to tell which ones are good and which ones are not so good until you get to interviewing them. So the whole game of hiring people is going to change dramatically. you will have to hire AI talent. It's almost unavoidable, especially on companies that are beyond very small companies, because these people have to manage your AI systems. Your licenses, your infrastructure, your open source platforms, deploy them, manage them, and so on. This is a job that doesn't exist in many companies today. Right now it's an it task. So maybe you can train some of your it people to do that, but that's a talent gap that most companies have right now. And then the last thing, which is scary, but connects to the background that I gave you before is digital labor is inevitable. It's coming. We will have significant parts of the business done by AI. What does that mean for your business? the different departments for your industry is something you have to figure out for yourself, but the sooner you start planning for that and thinking about it and talking about this with your leadership team, the sooner you can a prep yourself and be potentially gain huge benefits and market share because you'll be first movers in this particular aspect, finance and budgeting every, so there's, I've participated in multiple conversations about budgeting in companies, and as I mentioned, in a Fortune 2000 conference, it goes down to the following things. First of all, every department needs an AI budget inside the department's budget. Now this could be the department needs to assign a part of its existing budget to AI, which will force them to implement AI efficiently to gain this back. I really like this approach. but the other side is saying, it's not fully done. So I'm going to give you additional budget for AI implementation per department. In addition, you need a broader budget for the company to cover for things like additional infrastructure. So new IT infrastructure you don't have right now for AI applications. So whether it is simple as ChachiPT or more sophisticated, dedicated custom stuff, Initially for testing, but then for deployment budget for training, you need, we talked a lot about training. I'm not going to dive into this, but initial training and then ongoing training is something that's going to cost money and requires budget talent acquisition and retention. Is a big one and then inference, budget and optimization running these tools once you have them is not free, especially if you're running them against the API. I'll give you a simple example that I've learned from AT& T from a conference two years ago. AT& T currently right now spends 1 billion tokens every single day, 1 billion tokens every single day. So even if you're running on a cheap model that costs you 5 cents per token, that's 50 million. Her day, which means it's not doable. Meaning you have to find internal tools that you can run, like open source that runs on your servers. And then you're just paying for server time versus for tokens to somebody else. But even if we scale this down to a small business or a medium business, the cost of running these tools grows very fast as you add more and more tools. And yes, it's going to make. Everything else more efficient, but you can optimize for that. You can use cheaper tools for specific things. You don't have to go to Claude Opus for everything you need that costs an arm and a leg. You can find an optimize smaller tools, cheaper tools. In house tools, open source stuff, and a mix of those will allow you to optimize for the right level of expense to the right task over time while measuring ROI of these different things. Talking about ROI, four buckets of ROI. So everybody thinks, Oh, there's the efficiencies. There's a lot more than that. If you look at the bigger picture, and that's why AI is so important from a strategic perspective. One, valuation. If you can prove to anybody who wants to buy your company, invest in your company, and so on, that you truly are implementing AI and truly gaining efficiencies, you will increase the valuation of your company. That's huge as leadership or ownership people of companies. Retention of people. People are looking for this right now. People are looking for the next thing and AI creates excitement. And so doing that is huge. Capabilities building. So you can build new capabilities. We talked about this innovation that drive new revenue that didn't exist before that can drive ROI. And then the last one is traditional hard ROI, basically getting quick wins, getting efficiencies and so on. Consider all of those when you're thinking about AI implementation, AI use cases. the way to do this is to define initial POCs to define small projects that you can implement quickly to get quick wins and creating value. Try to create value for your clients first, because that immediately will generate higher ROI versus just initial savings in your company and then move to internal use cases. Moving the POC, the proof of concept to production is not trivial. There's a lot of roadblocks. So the fact that you've seen an immediate success with something doesn't mean it's going to be easy to deploy it company wide. And it's something you're going to learn over time on how to do better. You got to consider data needs. And you've got to start with initial consumer tools and start ugly. Just start with simple use cases that take into consideration data risks and try things out to get buy in from leadership. And if you're one of the leaders to prove to yourself that this is actually worthwhile and then move from there and move forward to most critical factors of AI success based on my experience and everything that I've read again, across the biggest data sources in the world. One. And none of them, by the way, is technology, which is not surprising to me, but might be surprising to some. One, the first one is leadership buy in. If leadership is not involved in the process, it is going to fail. So if you're the business leader, be involved, lead by example. If you're not, get sponsorship from somebody in leadership to be behind whatever initiative you're trying to drive with AI. And then the second one is training and education. AI basics, role specific, so use case training, and then continuous education across time. So these are the two most critical things. So quick summary. This was, I know a little bit Drinking from a fire hose, but AI is coming. It's here. It's just growing and moving very fast as we move forward. And you cannot just ignore this because your competition is not ignoring this. And if they will figure out a way on how to do this before you, you are going to lose market share to them because they'll be able to run faster, provide better service and products and do it cheaper than you, which usually means just one thing. As somebody wrote on LinkedIn, you have to get started and to get started, you need knowledge to gain knowledge. You can do several different things. You can figure it out on your own. Over time, you can listen to podcasts like this one. So the leveraging AI podcast or any other podcast, you can follow the right people on LinkedIn and different platforms and so on, or you can sign up for courses which dramatically accelerate the process, especially if you're looking for business specific knowledge. As I mentioned, the AI business transformation course, the next. the next cohort starts on October 28th. If it's something you want to do it, I would highly recommend that again, four sessions, four weeks. You will know before the end of the year, significantly more than right now. And you'll be able to start 2025 with the right foot forward. I really appreciate all of you who has joined us today. I know this has been long. I know this has been very intensive. But it's a very important topic. It could literally dramatically change the trajectory of your business for either good or bad, depending on the direction you need to take. And that's why it was very important for me to do this session now to help you guys prep. As I mentioned, don't wait for 2025. Even this was like a project. Prepare for 2025 session. You can start right now. So I want to thank all of you that I joined us on LinkedIn. I want to say thank all of you was joined us on the zoom. There's a few dozens of people on each side. I know you can do other stuff with your time and I really appreciate that you've decided to spend it with me in this hour. Thank you so much and I will see you next Thursday.

People on this episode

Head Shot

Isar Meitis

Host