The Truth About Off-The-Shelf AI

So there's lots of AI tools out there and they're meant for general market and the problem is when you have AI tools it is just a tool but when you have the off-the-shelf solutions that are AI driven you're not really going to get the best outcome from it because it's not really designed about you and your problems and your current your current strategy and this is the thing where I've been talking with quite a few clients and they said that oh yeah we've used AI before, it hasn't worked for us and then I'd ask them all what have you actually been doing and they say that they were using this very cheap software that's meant for general market but they're not generally, they're specialized, well of course that's not going to work. I've done it up in the cheap, you get the cheap results. Exactly and AI is only as good as well one the goal that you give it, the constraints that you give it, the instructions that you give it but also the data and a lot of the time when people put in the wrong data or poor data you're going to get poor results. So when I go into a company and I look at exactly what they're trying to achieve I understand their full process from start to finish. I know what the data is on the input, I know all the instructions that the AI agent needs to adhere by and then the output. You can't get that in off-the-shelf software and that's why they get off-the-shelf results. Yeah, interesting. I'm Mike Harris and I help businesses to successfully grow their turnover and profit and I work with small, medium and large businesses including startups and multinationals. Arguably the biggest challenge facing business leaders in 2026's AI but what does that actually mean to you and what can you do about it? So to answer these questions and a lot more, today I'm going to tackle AI head-on. And joining me to do that is a very special guest AI an automation expert and entrepreneur, Oli Wood. Welcome Oli, thank you for joining us. Tell them about yourself, your background and a little bit about AI in the business that you have. Yeah, absolutely. So my name is Oli Wood and I have a decade in tech. I've been in tech industry for all of my professional life and in the last couple of years I've specialised more in the AI space. I came from a data and automation engineer and then moved into AI automation and that's what I've been really focusing on and in terms of my businesses that I have, I've been working alongside litigation, so litigation analytics where we have a system where attorneys can upload all their case documents, fines, discrepancies and witness testimony and that's just one aspect of the many different companies that I've been working on. And that transformation for automation to AI, was that natural? Did you find it was all? Yeah, absolutely. Actually it was a step-by-step process. So what's funny is when I was about 16 years old, a lot of my friends had blackberries and one of the things that I noticed was if you texted blackberries enough, it wouldn't know how to handle so many incoming tech messages. So we just turn off and turn back on again. And then I was like, "Ah, that's interesting. I wonder if I can automate this." So when I was 16 years old, I thought, "How do I create an app that can do this automatically?" So that was my first ever introduction into building something that would automate something out of a harmless prank. And from then I always thought, "Well, how do I automate all of my life? How do I automate problems?" And from there I built up knowledge around automation and then AI came out and I'm like, "This is cool." Because if you have automation, if you have a system that you've really created that uses automation, if you sprinkle a bit of AI on top, you can do things that you've never thought were possible. And that's why I love this. And we'll come back to that a bit later, I'm sure we'll want to talk it through. Okay, well coming on to the podcast itself then, what part do you think AI plays in leadership? So one of the other things that you can use AI for and that is decision-making, there's only so much information out there when you come to a decision before you have to make it. And one of the things that I've actually used AI for is to help me decide what I need to do. Because since I've quit my job and gone full on in the entrepreneurship space, there's a lot that I don't know. So what I've been able to do is I've been able to use AI to help me make decisions. I mean obviously I don't know everything at the same time, but there are some decisions that it can make on my- -That's some of the gaps in it is. -Exactly. Exactly. So that's how I've used AI in my own life, in my own decision-making, to then move forward. And in leadership, that's exactly the same thing, because in leadership you need to make decisions. So whether or not it's a decision for yourself or for a team, that's how I've used it. Yeah, I think the other thing I'm just thinking about now, but the other thing about quite often when you're a leader, it can be quite lonely that you're trying to make decisions, see where you're going, you don't necessarily want to share what you're thinking with other people until you've finalized and completed those thoughts. And when I think about it now, how am I using AI? I'm using AI as to bounce ideas and just put something in there and see what can, you know, I mean generally it doesn't tell you and come back and tell you if it's a crap idea, but it can kind of see by what it's come back with that maybe it is just by the nature of what it's saying. Funny enough, you actually can get it to say that. Oh, can you? Yes. Yes. Yeah, so I've built some AI agents coaching AI. Well, suddenly the ones that are readily available don't tend to be. By default, by default, this is what I was talking about earlier with a very agreeable one, they're meant to be nice because it's meant for the general public. But some people don't want to have the general responses. They want the AI to actually just cut through the noise and tell them directly exactly what they should be doing. By default, all AI models out there at the moment, they don't do that. You need to make sure that your prompts, so the prompts that are behind it actually have better instructions that cut through that noise. So would it ask me the question at the end, you know, I don't know, I'm thinking about doing this. Would that be a good idea? Would that work with the NIA? I ask you that kind of question. Absolutely. Yeah, absolutely. Yeah, so that's an idea for a person then, is it, you know, it's just getting that make sure you don't get, yes, persons response, you get a proper, exactly. You don't think it responds, yeah. And you can actually say that, you can tell it, do not agree with me. I want you to challenge my thoughts. What am I assuming that I shouldn't be? Is it anything that I'm missing? And it will give you that response? And then you're like, oh, yeah, I forgot about that. And then that's where the idea generation actually gets a lot better. This is fascinating. Really, really good. I'm sure everybody who's watching is going to find this really interesting as well. So what do you think of some of the common challenges that businesses are facing at the moment with regards to AI? And as a second part to that question, what maybe are some of the solutions that will overcome those challenges? So there's lots of AI tools out there and they're meant for general market. And the problem is when you have AI tools, it is just a tool. But when you have the off-the-shelf solutions that are AI-driven, you're not really going to get the best outcome from it because it's not really designed around you and your problems and your current strategy. And this is a thing where I've been talking with quite a few clients and they said, oh, yeah, we've used AI before, hasn't worked for us. And then I'd ask them, oh, what have you actually been doing? And they say that they were using this like very cheap software that's meant for general market, but they're not general. They're specialized. Well, of course, that's not going to work. You're going to get results. Exactly. And AI is only as good as, well, one, the goal that you give it, the constraints that you give it, the instructions that you give it, but also the data. And a lot of the time when people put in the wrong data or poor data, you're going to get poor results. So when I go into a company and I look at exactly what they're trying to achieve, I understand their full process from start to finish. I know what the data is on the input. I know all the instructions that the AI agent needs to adhere by and then the output. You can't get that in off the shelf software. And that's why they get off the shelf results. Yeah, interesting. So what are some action all tips that you think you could give to our viewers and listeners who are regarding AI? Stop using it for Google. It's my first one. All for creating co-catures of yourself that you're supposed to do. That's just you. I mean, you can't really widen me up at them, I'm just saying that. I just think there's this brilliant technology and everybody seems to be just using it. I'm sure they're not, but that's very good. Actually, that is a very good point. I don't like how a lot of either marketing agencies or companies that are doing marketing, they're using AI generated images. Although I put a caveat on that, I have worked with a couple of physical product brands. So they do like skincare and they've given me a couple of images. I've created a couple of very professional images of their products. I've created AI generated images from that in scenes that are they look very good. Then what they do is then they put the graphics on top so the words and stuff like that. Those are really good because that has a purpose. It's not just generate me a product that looks like that. Those images really annoy me because they don't look great. They look clearly AI generated. But you need to be a creative director and then you can really create. That's your target isn't it? Totally creative director. When it comes to AI generated images and AI generated videos, you need to make sure that you are creative director. You can't just have like any body off the street can create something that looks great and a lot of people will be fooled in terms of they don't think it's AI. You need to make sure that the people who are creating these images and these videos, they have the background in image creation and also videography before they even touch AI. If you've never done that before, don't do it because it's quite clear that it is AI generated. So that's a big don't. Yes, so back to tips, the actual tips for business. There's a lot that you can automate in your business and what I would highly recommend doing is never using AI to try and automate a process that doesn't work because if you're going to get AI to even touch a process that clearly isn't giving you the right ROI, why do you think AI is going to fix it? You need to actually go back to the drawing board. You need to make sure that the process that you currently have actually works because I've worked with companies before or actually I haven't worked with these companies before because they didn't have a process that worked. So for example, when I was doing lead revival, so that was when I was creating AI agents specifically for to reaching out to old customers and then booking them into a sales call automatically all on autopilot. Some of the companies didn't even have the right processes. So for example, if I were to have them book into a call, they didn't have the sales team ready. It's like, well, okay, well, I'm not going to work with you then because you haven't even got the basics right. That's the thing. So whenever I create something, creating a AI agent for a company, I need to make sure that their current process works because I'm not going to fix that for them. I'm just going to elevate exactly, sorry, I'm going to magnify the process. Yeah, the problem. So if they have a problem, I'm going to magnify it. And they'll be like, oh, AI is awful. It doesn't work. Like, no, your process didn't work. Yeah. But as you say, coming back to your example, if they haven't got a sales team to be able to deal with the calls or the clients when it's gone through that mechanism, it's going to fall over. Well, it's going to fall over where they bring AI into it anyway because there's nothing there to pick up on the on the pipeline as it comes through. Yep, exactly. Any other tips? Yes. So another one being, we've mentioned it before, but giving it more context than it needs. Because I still don't even think that when I'm going through all of my problems that I'm really trying to overcome, I still don't even think I give it enough context. I mean, yeah, it gets me to the end goal, but it could have got there soon, if I gave it even more context. But also whenever you talk with AI and it's a very hard problem, when we talked about before, give me 10 questions to try and get to where you need to be. And I think that is it. Those are the biggest tips that I would suggest. Also breaking down your problems into digestible chunks would be a good one to do. Because I have worked with some companies and they say, oh, I want AI to do our marketing strategy and then, oh, I also want it to talk to all of our leads at the same time. Oh, and then I also want it to do my admin. And then, oh, actually, I also want it to do my hiring process. I might not. Okay. But why don't we just tackle one thing? One of those, yeah. Because you can't buy all the whole ocean at the same time. That's not going to work. So. That's no dim to anything in business. If you've got too many initiatives going on at the same time, it's never going to work. And then you've got to, and then guessing AI is exactly the same. You've got to hone me in on the priorities and then have you say one day at a time, just one day at a time, just one AI agent at a time tackling one problem at a time. Because you can't get an AI agent to do everything. Because AI agents are very good at solving a specific task given specific instructions. Just like people. Exactly. That's why we have different department heads. Because they focus on very one specific thing. So if you have an AI agent that does everything, where you're just going to get generalized results, actually probably even worse in generalized results. Because it's not great at that. And this is why you take every single individual problem that you have, break it down into very small chunks. You might even need more than one AI agent for one problem. And in fact, most of the problems that I solve, it's not using one AI agent. Because I know processes, I know automation, and I can break everything down into very simple problems. And I just do that over at a large scale for one specific goal. That's how I do it. So what you don't want is that old phrase, "Jackable, trace master, none." Yeah, a plaster AI. That's the only thing that doesn't want the master of one. Yeah. Over all of your problems. Interesting. Yeah. What about some AI do's and don'ts? What would you do? And I think maybe you've covered this in some part already. But yeah, just a few do's and don'ts with that. Well, definitely not maths. Never use it. Just use a calculator. That's what it's designed for. Do's, giving it more context than you think it needs. Because you probably think that you're giving enough context. You definitely know. If you have a problem, I would certainly try and use AI to understand, at least from an idea bouncing point of view. Because there's probably a lot of misconceptions, not so misconceptions, but there's probably a lot of ideas that you're missing out on because you haven't challenged yourself. You can use AI to challenge you and your own assumptions on your problems. And that's what I use. And the reason why I'm suggesting that this is a definite do for people is because that's exactly what I've been doing for the last six months. For the last six months, I've learned more doing this in the last six months than I have in the last 10 years because I just assume I know nothing and I want it to teach me how to do things I've never done before. Like explain like I'm five. Give me this step-by-step instructions. Okay, but why did you do it? Why do I need to do this instruction? And then it takes me down this massive route and I'm like, "Oh, okay, so that's why I should do that." And I've learnt so much doing that. I've never been able to do in a classroom. And actually, this is funny. One of the other things that I really want to get involved in and I believe that there is a massive gold mine opportunity here and it's the education space. Education, you can completely revolutionise it if you had individual AI agents that know how you should be taught. The best knowledge retention strategies that an AI could give you, I believe that's a massive thing for the education market. No one's really doing it to my level. That's right. Yeah, maybe the. Maybe the action. Yeah. Well, before we post this up on YouTube, I'll give you a few days ago and I'm peeing to the idea. Okay. Now, and I think you're right because I mean partly because it would revolutionise the education. But children coming in, it's like everything, isn't it? For me, this is entirely new. To my 10-year-old, she's seeing this as new. It's not my 10-year-old, I'm a grandchild. And then, but to my other grandchild who's six, he's all over AI. So even that difference of four years is a difference to how they see AI. So the children are coming into education now. They're the ones. It won't be the ones at high school. It'll be the ones who are coming into primary school now. That's where AI could. Because they're living without any way. You know? Yeah, they are. Fascinating. I know that they are trying to use it. Well, I know that teachers are using it in education. They're using it to, well, one create lesson plans, which is fine. But they're also using it to mark homework, which. Not a massive fan of, just because if you're over relying on AI and you haven't designed it correctly, it's going to give like poorer results. Isn't that cheating and reverse? I mean, if it's not right. If it's not right for the student to use AI to come up with the answers for the question, isn't scoring it using AI exactly the same mistakes? Exactly. Exactly. It's a bit hypocritical if teachers say, "Don't use AI for your homework," but then they use AI to mark it. There's an imbalance there. It does see if it's. This is why it's not a simple problem in the education. I have theories on how it would work. It doesn't actually mean it. But it certainly worked for me. I genuinely believe that if everyone had an AI tutor in their own pocket, they'd be much much more than they are now. I guess I think I'm doing this already. AI is really replacing search engines to some extent in all. Okay. You think? Yes, but it shouldn't be used that way. No. I can't go. Another mid-debunt. So, if you're using AI. We should have a sounder time be that goes off from time when something else should. Don't use AI as a search engine. Right. Okay. There was this. There was a statistic that I saw fairly recently, which was the average length of a search that a boomer does on Google is longer than the average input that's most like millennials put in AI. So if 60-year-old is putting in a message into Google that is longer than the average message length in AI, there is something going wrong because this is not how you should be using AI. Well, I'll put exactly the same message into Google or into AI. Okay, well if that's the case then just use Google. Yeah. Because AI is not a search engine. I know that open AI is coming up with their own stuff. That's designed slightly differently but even then just use Google. Okay. The mother myth to my trouble site. Yeah, interesting. Yeah. Okay, well I think that probably closes for today talking about AI. I have to say I have found it really fascinating. I've learned a lot and what's one of the reasons I do these shows is because it allows me to go off and think differently and do things differently. So today's been absolutely brilliant. So thanks for coming on. I always like to finish the show by setting out audience at some homework that they can go off and think about and implement some of the things that come out of them today. So I thought I'd ask you if you'd like to set today's homework. So one of the first on the list is don't treat AI like a search engine. We've already talked about that. That's two, there's three reasons why but AI shouldn't be used as a search engine so don't use it as one. So one of the other things that you can also do is write down every single decision that you make. In fact, actually just write the top 10. Write down the top 10 decisions that you make on a daily basis. Write down your exact steps and then pass up through AI to audit and understand if this is the right strategy for you. A lot of the things that I've done in my past in the last six months is I would always give AI the exact context of why I come up with these decisions and then get it to challenge me and my assumptions. And I've learned a lot from that and it's given me better advice and I've always adapted based on the information that's provided me and if you're not already doing that I highly recommend that you do because you have no idea if the current process that you're doing is the right one and you wouldn't know unless you asked. Another thing that you can do is look into your own business. Ask yourself what resource, what leads are just just lying there that you could actually convert into new business using AI and automation. I've built a company around reviving old leads and most companies right now have a database full of old customers or customers that or potential customers that have inquired before but they never actually converted. Ask yourself can you use that database of old leads to then convert into new revenue using AI and automation. So when you write your next prompt there are six steps that you need to ensure that you have to make sure that you're going to get the output that you need. So the first thing you need to give the AI is it's exact role that it needs to play. So is it a marketing director? Is it a strategist? Like you need to be very specific in the type of role that it needs to achieve your goal and the next part is the objective. You need to give it the end goal and be very specific in what it needs to achieve. Otherwise if you aren't clear on that why do you think it's going to give you a great output? And the third one is context. You need to give it all of the data that it's required to be used to get to your end goal because if you don't give it all of its context it's going to make decisions or outputs based on imperfect data. And then you'll use that and it won't work for you. So you need to make sure that you've given it all of the context. And then the fourth one is you need to give it the constraints. What can it do or what can't it do to make sure that you are getting the right result based on your outputs that you need. And the final one is the output format. So it might be a summary, it might be a full document with very specific headings. You need to be very specific on its output. All of those points you need to make sure that are within the prompt to ensure that you will get the best output for you. And the last one is when you have a very hard problem to solve you need to spend at least 20 maybe even 30 minutes at a time dictating to AI your exact problem giving all of the constraints, giving all of the data before you even think about executing. Because once you have that full plan then it's much easier to achieve what you're trying to do. Excellent. What you said particularly that last one is just what I would advise when you're doing a strategy generally. Think it through, set it out. Look at all the things that you need to consider. Explain it properly before you even get into sort of sharing it with people and the implementation. So fantastic what you've just said. And I'm as I said earlier I've learned a lot from this and I'm going to take what you've just said into a crypt sheet. Although I've got to say that last one I'm preparing for 30 minutes before I actually feed it into AI. 30 seconds as well. You need to up those. That's a permanent exaggeration. I'm certainly not under 30 minutes. So that's really going to be a challenge for me to do that. But absolutely. So once again, thank you for joining us, Oli. It's been absolutely fantastic. Hopefully you'll come back and join us again for a future episode. I know the audience are going to be all over this and going to have loads of questions. So we'll pick that up on a future. I'll maybe get you back to answer some of the questions. Thank you for joining us today. I hope you've enjoyed it. And you come back again soon when we'll have another guest providing some excellent top advice for business leaders. I'll see you soon.