What if AI could 10x your business growth?
On today's episode of Leveraging AI, Isar Meitis talks about the secrets to leveraging AI to take your business from surviving to thriving.
Whether you're a startup or an enterprise, these practical AI strategies will help you tap into astounding growth opportunities.
AI news of the week:
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!
Hello and welcome to Leveraging AI. This is Isar Meitis, your host, and in today's show, we are going to cover four concepts you must know in order to maximize the ROI of the AI efforts in your business. At the end of the episode, I am going to share some big news from the AI world that happened this week. But now let's dive into the four concepts that can help you to maximize the ROI of AI implementation. Thanks Hello, and welcome to Leveraging AI. This is Isar Meitis, your host, and I have a special episode for you today. In the past few months, I had the opportunity to speak in multiple conferences and teach courses and consult to companies and in general have multiple conversations with hundreds of business leaders and managers around the topic of how to successfully implement AI in businesses And I came up with four concepts that, if you know them, will enable you to dramatically increase the ROI of the implementation of AI in your business compared to if you do not know these concepts. And since I started talking about them in speaking gigs that I do in large conferences and getting a very good feedback on them, I decided to share them here on the podcast. The first concept is stop thinking efficiency, start thinking outcome. What the hell does that mean? It means that as business owners, we are trained to think on how to get small incremental efficiency gains in different processes that we do. Why? Because this was the only way to get better before AI was available. Meaning we're looking how to gain 5%, 7%, 8%, 10 percent efficiency on various aspects and processes, usually roughly keeping the process that we have always done. And now with AI, This is not the right way to approach this. And what I mean by that is that you need to look what you're actually trying to achieve, because it might be that with AI, you could circumvent 2, different steps in the process that you're trying to do. So instead of gaining a 5 percent efficiency in a step, you can go straight to the end or save several steps on the way and saving 70, 80, 90 percent of the process instead of getting a 5 percent efficiency gain, like we used to in the past. So let's give a few examples that will hopefully make this a lot clearer. Example number one, customer service. What is the goal of customer service? The goal of customer service is not to solve tickets faster. It's not to have a good ticketing system. It is not to answer the phone faster. It is not time to resolution. The goal of customer service is to have happy customers. And again, if you think about systems and processes, we knew before AI, well, it touches all the things that I said before, it's having more people in the contact center, is being able to better analyze and assign the right tickets to the right people so they can resolve faster. It's all these little things. But today with AI, you can have an AI based contact center that can answer the phone 24/7, 365 days a year, be connected to your entire data set so people don't have to transfer and jump between five different departments in order to get all the answers and he will do it at a fraction of the cost. So not just the customers win, but you win as well. So instead of gaining small benefits in solving tickets faster or answering the phone faster, you can go to having happy customers at a much higher success rate at a much lower cost. And that circumvents many steps along the way that we used to think about as mandatory and necessary. Example number two, writing proposals. Writing proposals, depending on where you are, depending on which industry you are, what kind of customers you serve could be a very long and tedious process. It requires doing a lot of analysis of data, whether it's customer specific data or cost data from your side. Combine them all together, understanding exactly the goals of the project or the service that you're trying to provide. And then writing a detailed proposal. Well, now you can do this in very few steps in a system like ChatGPT which is either free or cost 20 bucks a month, which is again, in a business perspective, free. And to do that you can take the data load it to advanced data analysis module in ChatGPT, have it do a lot of the data for you upload additional information like PDFs and images and drawings and charts and graphs and any information you need in order for ChatGPT to understand exactly what you're trying to do and have it either write the full proposal for you or write a very good initial draft. So it goes from I need to now, first of all, spend three to four days or a different team has to spend three to four days to analyze data, to send it to a salesperson so then can write the proposal to then be reviewed by whatever manager. Many of these steps can happen within minutes instead of days done by a single person. Another great example is. SEO. Historically, SEO required multiple different steps. And it's still the case today, but each and every one of those steps, literally almost each and every one of them from keyword research to on page analysis to off page analysis to. Writing the keywords that you want to target to creating the draft content, writing the actual content, each and every one of those steps can be dramatically assisted by AI, either a dedicated tool that is built to do SEO or write content, or even in ChatGPT or Claude or Bard or one of these tools. So. going back to the initial concept, it's not a 5 percent increment. It could be a 50, 60, 70 percent savings in the time you have to invest while getting the same or even better results compared to using traditional ways. Two additional examples. if you want to learn more about AI based customer service and, or even sales contact center, you should check out episode 36 of this podcast In which I interviewed James Lindsay and we go in depth into this topic and the capabilities that exist today. Another great example of what you can do with AI in order to save dramatic time is planning and executing local events. You heard me say that many times before on this podcast, that connecting with people, the human connections that you have are going to become a lot more important than they are today. And they are very important today. And the reason they're going to become so important is because with AI, everything else, everybody will be able to do pretty damn well. And so the differentiator that you're going to have is the relationships you will have with clients and prospects and other people in your ecosystem. And so running local events becomes a huge benefit to companies in order to connect with relevant people. But designing and creating and running events is a huge task, or at least it used to be. It used to take weeks to plan such event. And now you can do it in minutes or a few hours, which makes it an extremely high ROI. So how do you do that? Well, go and check out episode 31 of this podcast, where we talk to Ashley Gross. And within a single hour, she took us through the entire process that she follows in order to run such events. Locally, going back to what I said in the beginning. A process that used to take weeks now takes minutes, and maybe it's not 60 minutes, but 90 minutes, but it's still a savings of probably 95 percent of the time it used to run the same process. And the last example I want to give you has to do with creating videos for either training or marketing. If you want to have a face of a person, an avatar that will represent your business. In a video that is explaining something that could be how to use your product. It could be a customer service thing. It could be for internal training, external training. It could be a marketing thing. You can have a virtually created avatar that is either made up or looks like a real person such as you or one of your salespeople, marketing people, and so on. And you can create such videos in minutes Using tools like Synthesia or Haygen or DID and in these tools, you can create a script using AI and then have the avatar say that thing while having a presentation running in the background, which can also be created by AI. So taking video production that used to take weeks and cost tens of thousands of dollars while hiring actors and voiceover people and videographers and lighting people and sound people and editing and all of that literally goes away and you can do this in minutes. And if you want to learn more about this topic and where the video production world is heading, when it comes to business related videos, you can check out episode 14 of this podcast, where we dive into this topic with Yuval Makhlin. So quick summary of concept number one, Stop thinking the way we are trained to think as business people, which is looking for small incremental efficiencies. Think of the outcome and check if there is an AI tool that can take you there in significantly fewer steps than you're doing today or dramatically shortening the steps or at least one big step in the process. Meaning instead of getting a five to 10 percent efficiency, you can get a 95 percent efficiency while maintaining the standards. And in some cases, even getting better results. The second concept is also very important and I call it from profession to skill. Back in the 17th century, there was a profession called computer. And that profession was people who knew how to compute. It was people who knew how to calculate stuff because there was no other way to do this. Well, today, I don't know anybody who has computer written on his LinkedIn profile, and probably he, you don't know anybody like this either. Same thing as typist. When I was a kid, our next door neighbor was a typist and a lot of other people were typist. And again, most people today don't use typist anymore because most of us learned how to type pretty well, because it became a part of what we have to do. And it became easier to actually do that. So it supports the way we do business. So what does that have to do with AI in your business? Well, AI enables a lot of things that used to be professions to become skills that anybody can do without a huge learning endeavor. As example, graphic design. You can create incredible images and graphics for any need that you have, Whether it's your website proposal, marketing needs, et cetera, you can create amazing graphics without a graphic designer, just by learning how to use tools like Midjourney and Dall.e 3 and Ideogram and many others that can generate images and graphics right now. And it doesn't mean you don't need a graphic designer. It just means you need a lot less graphic design time in order to achieve similar things that you did before. And it also means that people that are not graphic designer can have the skills that a graphic designer has and get 90 percent of the result that a really fully trained graphic designer would get to. And doing that, once you learn how to use the tools in minutes, instead of hours. Similar concept goes to code writing. So do I think that people who never learned how to be software engineers can now become software engineers overnight as a skill? No. But can you start writing small pieces of code or understanding pieces of code without having any skills? Absolutely. I do this now regularly. I create extensions and plugins for Google Sheets, as an example, and I've never learned how to write a line of code ever in my life. Does that make me an expert software engineer? No. But does that enable me to do a lot more with writing code than I've ever done before, which was zero? Yes. Absolutely. Yes. And again, it's a skill that anybody can learn. And the same concept goes to being an accountant Or writing scripts for plays or TV shows or marketing videos. The same concept goes to being a paralegal and so on and so forth. So many aspects that relate to a company running efficiently that we used to think of as I need to hire a person that has this profession is not a necessity anymore because it's a skill that people within your business can learn, or if you want to twist this around, when you do your next cycle of hiring, you have to think about what skills You need in your business and not necessarily what professions you need in your business and hire people who have the skills to use AI tools, to achieve those goals that you're trying to achieve. And that does not mean that they necessarily have to have A degree or a professional course that they've taken on that particular topic. It's a completely changing mindset that can help you run your business with the existing people much more efficiently reducing overhead that you're currently contracting externally, and definitely should impact the way you hire moving forward. Concept number three has to do with two ways that you can win in the AI era. And these two ways is either proprietary data or optimization. And let me elaborate on this concept. If you have proprietary data and most companies do, you can train models on that proprietary data, which means you can now have a way to ask questions, get information, provide value. Based on data that you already have. And you can either use this internally to make better decisions, to support customer service people, to support proposal writing, to support your accounting process, or you can use this externally as well to give straight access to this, to generate quotes on your behalf or to provide better customer service, et cetera. Now, many companies think, well, I don't have any proprietary data. I'm not Amazon. I'm not Google. I'm not Amdocs. I'm not Lockheed Martin, et cetera. The truth is, almost every business, if not any business, has proprietary data. You have all the proposals that you have written. You have all the project data that you have accumulated with time tracking of different people in different departments, etc. Many businesses have recordings of their sales calls, recording of their customer service calls, chats, open tickets, and so on. Each and every one of these is a data point that you can use in order to train an AI model in order to give you more information when you try to do these things in the future, which will give you a very high ROI. So Option number one to win in the AI era is to use your proprietary data. And the reason it's unique is because nobody else has that data, meaning by definition, your specific bot that is tailored to your needs and the needs of your clients will be the only one in the world, or one of very few that has access to that information, which will enable you to really enjoy the accumulated knowledge that should exist in your organization in order to be more competitive in your field, in your niche. Now the other option is optimization. If any business that will learn how to use AI tools to do the day to day that they do anyway, using AI in a much more efficient way than their competition is going to win because if we combine the things we talked about before, you can have not just a 5 percent efficiency gain, but sometimes 70, 80, 90 percent efficiency gain, meaning you can reduce costs while still providing high value to your customers while having higher margins without really giving up on anything, which means everybody wins. Your clients get faster results, you're delivering the same or better value than you did before. You do this at a lower cost and you can make more money. So learning how to use AI tools effectively can make a very big difference between you and your competition. If you can combine these two aspects, proprietary data and optimization, you can win on both parameters of this equation, making you a lot more competitive and allowing you to gain even more market share compared to your competition. The last concept is also not easy to grasp, but once you see a few examples, it will be very obvious And that concept is that AI enables infinite scalability. I know what you're thinking. Well, there's no such thing as infinite scalability. And that is correct. There is no such thing as infinite scalability. But there are aspects of the business that with AI can scale way beyond anything you can imagine and way beyond the ability of other aspects of the business to scale, making it an infinite resource in that particular aspect of the business. As an example, your marketing team's ability to create social media content was limited to X number of pieces of content per day or types of content per day. And now that literally goes away. You can generate as many high quality marketing ads or organic types of content in a hundred X, the amount you could create it so far, or 50 X, which means you will probably won't be able to post it, but it also means that the stuff that you can post, if you focus on the right things, if you focus on the right message, if you focus on the right audience, analyzing that right audience, all these things can now be done at a much higher scale, dramatically reducing your dependency on the throughput of bottlenecks that every business has in its marketing process. I just gave an example earlier of creating marketing or training videos. The ability to do that costs a lot of money and took a very long time and now takes literally minutes. So you can now create as many of these videos as you want. Meaning for that particular aspect, infinite scalability. Another example is writing proposals. The actual writing of the proposal take a lot of time. The analysis of the information required takes a lot of time, which means you were limited to writing X number of proposals. So even if your marketing did an amazing job and got you a lot of candidates that are actually willing to buy from you, the next bottleneck becomes, well, I can't write proposals fast enough. Well, you actually can. You can take that to the HR world as well. How many job applications can you review in 20 minutes? Well, historically, if you're very good, probably three to four, maybe five. Now you can do the initial screening using AI basically a limited number of them just by loading them to advanced data analysis, telling it what you're looking for and having it analyzed for you and give you any statistics and any results and any filtering and any feedback on any candidate based on the applications that they have submitted. So really almost every aspect of every business has these. components that used to be bottlenecks, which now should not be bottlenecks anymore because for that particular step, you can achieve infinite scalability if you use the right tools and the right processes in your business. So let's summarize these four concepts and what do they actually mean to your business. So number one is stop thinking efficiency, start thinking outcome. Number two is from profession to scale. Number three is proprietary data or optimization as ways to win in the AI era. And number four is AI enables infinite scalability. Have those concepts in the back of your head with every new big decision you're going to make for your business Because it will enable you to have a completely different lens through which you evaluate your future decisions in your business. What people to hire? How many people to hire? What kind of processes you need to put in place? What kind of clients can and can't you approach? What are going to be the client's actual needs now that they have the same kind of concepts in their heads and so on? Use these concepts as another filter when you make future business decisions and you can gain much better results than people who won't in this new era of business with AI. I would love to hear your thoughts on these concepts. I would love to hear any questions you have about these concepts as it relates to your business. So please reach out to me on LinkedIn. It's Isar Meitis, I S A R M E I T I S And share what you think about this concept, share any feedback or any questions that you have. I would gladly engage with you there and provide you any answers or any feedback to thoughts that you have. In addition, if you like this episode, if you like this podcast in general, I would really appreciate it if you pull up your phone right now, assuming you're not driving and give us a five star rating on your favorite podcasting platform and recommend the podcast to other people that you know, that can benefit from it. This is your way to help us reach more people and help other people learn how to better implement AI to have more success in their businesses. I would really appreciate your support with that. And now to the news from this week. The biggest piece of news from this week comes from an interview that Sam Altman, the CEO of OpenAI, the company that gave us ChatGPT, has held with the Financial Times, where he's speaking about ChatGPT 5. So the next generation of the ChatGPT that we're using right now, which is 4, or if you want right now, it's 4 Turbo, but it's still based on the same model that they've trained a while back. He is claiming that the next version of GPT will basically achieve AGI or something very, very close to that, meaning it will be able to do everything a human does, not on a physical level, obviously, but on any other level at or beyond human level. He also said that it will take billions and anything between months and years before they can complete training this model. He was very supportive of their partnership with Microsoft. And he's clearly stating that this partnership is helping both sides on open AI is obviously providing them access to computing power as well as 10 billion in budgeting. But also, as he said, and I'm quoting Microsoft has shifted its entire business model around the use of AI embedding its copilot system built on top of GPT-4 into Windows, Microsoft 365 and other products. The take from this is that open A. I. S. Actively working As expected on achieving AGI artificial general intelligence. and in order to do that, they're using the huge funding they got from Microsoft. Nothing too surprising there, but it's a confirmation that this is what they're working on. And this is what they're striding for. He even said, you know, GPT 3. 5 was just text GPT 4 can now do all these other things like understand images and create images and create code. And speak and understand natural language through voice. And the next iteration is going to be doing basically everything humans can do at or above human level. And speaking about humans and human behaviors and the impact of the AI world and developments in technology on that, China just disclosed plans to I quote mass produced advanced level humanoid robots by 2025. That's basically a year and a quarter out where they're saying they're going to humanoid bots that are disruptive to the way we live our lives today, and they can quote reshape the world. Now, they are not the first group that is working on that. There are companies here in the U S like agility, robotics, and Tesla an Amazon that have been working on such robots for multiple purposes, but they're the first one that are saying that they're going to mass produce them in such a short period of time. Now, how much can we trust these claims? I don't know, you know, they come from China and they made some bold claims previously that did not materialize, but they're definitely working in that direction. And let's say they get it wrong and it's not 2025, but 2026, it still will have dramatic implications on how we perform tasks. And now it's not going to be just white collar tasks. Like we're talking with ChatGPT and large language models, but also. Physical tasks that can be now performed by these bots. Meaning many, many jobs in addition to white collar jobs are at risk within the next few years from this technology. And the flip side of that with a more optimistic approach came from Elon Musk this week that at the UK's AI summit said that AI will inevitably eliminate the need for human jobs entirely. And he was talking about a future of abundance, and saying that universal high income is very likely now a lot of people have been talking about universal basic income or UBI as something we should stride for with the use of AI. But a lot of people on the other hand, send that basic income is not what the humans want. And this is the first time I heard somebody, especially at the level of Elon Musk saying that universal high income can be achieved. With the use of AI, how does a society work that way? I don't think anybody has a clue, but the fact that people are investing time, money, and resources to getting us there it's somewhat comforting, but that being said, there is a period between when we achieve either universal basic income or universal high income and the time we have right now where AI is going to take or eliminate a lot of jobs. And to be honest, I don't really know what that entails for us as a society. But I assume it will be challenging. And in a similar topic of finding jobs, an unemployed person was using AI in order to apply for jobs. He has used AI in order to write and submit applications to over 5, 000 job openings. He's a software engineer and used a tool called LazyApply to submit these 5, 000 applications. And he managed to get 20 job interviews through this application and even got a job through that process. The good news in this process is now simple people like you and me can use AI in order, if you want to level the playing field against a very sophisticated and in many cases, tedious process of job hunting. The problem with that is that using tools like that can flood employers with low quality and inadequate applications, which in the long term will actually make the process of finding a job even more troublesome. The really interesting and important fact that came from the interview with this person is that the highest quality of job. Opportunities that he got came from his personal outreach through people that he knows. Now you heard me say that time and time again on this podcast, that human connections. I've always been a very important part of successful business, but it's going to become. A lot more important and the reason it's going to become a lot more important is because everybody will be able to do a lot of things at a good quality at scale because everybody will be able to use AI tools. And the thing that's going to differentiate you and or your business is the ability to build strong and long lasting human relationships. And speaking of humans and its relation to AI, an interesting research just came out that shows that adding emotional stimuli to your prompts when you use AI can improve the AI model performance, meaning usually positing words like confidence and success, trust, excitement, building them into your prompts will enhance the quality of the response. The same thing, by the way, in a different research found that being polite to the model will get you better results. So saying thank you and encouraging the success of previous steps. We'll get you better results in the model. While this sounds almost crazy because you're talking to a machine, you need to remember that machine was trained on huge amounts of human communication. And guess what? When you're nice, polite, and positive, when talking to other people, you get better results. So that's the standard, if you want the benchmark that these models learn from. So using this is not going to just help you talking to other humans, it's also going to help you get better and faster results from AI models. Another interesting piece of news that came out this week, that relates to human communications or not human communications anymore, is that a UK firm called luminance developed an AI system that autonomously negotiated a contract with another AI in this particular case, that contract was an NDA that it negotiated with another AI in this minutes. So what does that tell us? It tells us that the future of at least basic legal agreements between people and companies will be AI based because every side's AI will know what it's trying to achieve and it can very efficiently negotiate with the other AI and get to an outcome most likely much faster than any other human while taking into consideration The legal ramifications and needs based on where the company is, what the rules are, and so on. I see this as a very positive development from an efficiency of doing business, unless you are a lawyer that specializes in these kinds of things. Another big piece of news when it comes to new AI technology came from Samsung this past week. Samsung unveiled its own generative AI system and it is a suite of AI tools that are built into Samsung devices running on their phones, tablets, and so on, and it includes a large language model code writing assistance, image generation and editing, etc. built straight into phones. They are the first. Phone company that announces a full suite like that, but it's highly anticipated that Google and Apple with iPhone will follow this suit and will introduce a I tools that are baked straight into our handheld devices. And speaking of big developments and AI capabilities that are baked into existing capabilities, GitHub just launched Copilot Enterprise. So GitHub has launched its coding Copilot tool a few months back, that enables companies and individuals to create code using prompts, using the GitHub database as the training data, they've now introduced a new level that they call enterprise. And the biggest difference is that the enterprise level can look into your own company code as part of its training data, hence making the code that it writes connect better and be seamlessly integrated with the code that already exists that a company already saves on GitHub. This is obviously a huge benefit from a company perspective and it's obviously another step in the AI wars, in this case, between Microsoft, who owns GitHub and Amazon that owns Code Whisperer. And so we're going to continue seeing these innovations happening very, very fast in order to provide additional value and compete on market share in this AI code generation war, and now, three rapid fire updates on new innovative generative capabilities. Adobe and Lama had just introduced new capabilities that can generate 3D models very, very quickly. Adobe researchers develop a tool that can take a 2D image and develop a highly accurate, high definition 3D model of it. In five seconds. This is obviously a huge game changer for industries like gaming and design and augmented reality and so on, allowing the creation of 3D models, more or less on the fly. And the Lama model is a little different. It's not based on an image, but it's allowing you to create 3d models based on text prompts. So a different approach with a similar outcome. I think over time, we will see all of these converge that will allow all of these to work either one way or the other, or in a combination of uploading an image and then adding text and then getting the results. It's very exciting. And I've seen more and more of these 3d announcements. And I think That combined with video is probably the biggest things that are coming in the pipe in the beginning of 2024. We'll start seeing more and more tools come out that can do these kind of things. And with that. To speak about creating video runway has just introduced what they call motion brush, which is really, really cool. And if you haven't seen it, and I'll put the link in the notes, like the notes for everything else, you can take a static image and use that magic brush on any section of it in order to make that particular section of the image be animated in the video that they've released. They have a picture of a boat in a dark night. In the ocean and they use the brush first of all on the skies which makes the cloud start moving and then they use it on the boat and the boat starts riding in the ocean or they use it on a bus, a static bus in an image and the bus starts driving on a smoke from a cigarette and the smoke starts moving or water in the waterfall and so on and so forth. So you understand where I'm going with this. It basically allows you to animate in a very high level of detail just by moving a brush on top of a section of an image and this could be a big section. It could be a whole area of the image like water in the ocean and then the water will start moving It's almost magic, but it's available right now on the runway platform. That's it for this week. Continue to explore AI, try new things, share what you find with the world, share it with me on LinkedIn and until next time have an incredible week.