Integrating AI Into Business Strategy: What You Need to Know

AI is quickly reshaping the way organisations operate, from streamlining everyday tasks to enabling entirely new ways of working. Yet despite the excitement surrounding AI and what the future holds, many businesses are still grappling with a fundamental question: how can they adopt AI in a way that delivers real value while managing the risks?

• Connect with Rob on LinkedIn

• ramsac: award-winning technology and cybersecurity consultancy - https://www.ramsac.com

• Rob's latest AI project - https://thoughtprovoked.co.uk

• Explore Rob's collection of books on cybersecurity, leadership, AI, and technology  

Episode Transcript

Welcome back to The Transfer Files. Today, James and I are joined by a true authority at the intersection of two of the biggest challenges facing organisations today, artificial intelligence and cybersecurity. Our guest today, Rob May, is a globally recognised AI and cybersecurity thought leader, international keynote speaker, best-selling author, and a trusted advisor to businesses navigating both the opportunities and risks of emerging technology. He is the founder and Chairman of ramsac and serves as an ambassador for the Global Council for the Responsible Use of AI. With 35 plus years of experience in the industry, Rob has built a reputation for helping leadership teams understand how AI is transforming business. He's worked with organisations around the world, sharing insights on everything from AI ethics and responsible adoption of cyber resilience and building future-ready businesses. I am very excited to welcome Rob to the podcast today to explore how AI is reshaping the way businesses operate, the cybersecurity risks leaders need to be aware of, and what organisations can do to integrate AI safely. Let's get into the conversation. Rob, a very big welcome to the podcast today. How are you doing?

I'm really well, thank you. It's a pleasure to be here.

It's a pleasure to have you here. Thank you for making your time. Yeah, and welcome James, as always to the podcast. How are you doing?

I'm very good today, thank you.

Good, grand. Now, before we get into the depth of the conversation today, Rob, are you able to share a little bit about you and who you are and how you have landed in the world of AI today and what you're doing and what you're currently doing.

Okay, so I describe myself as a technologist, but 35 years ago I started a technology business called ramsac. And back in those days we were traditional IT services. And then as time progressed, we became more and more involved in cyber. And it was really through cyber that I got pulled into the world of AI. AI is such a big part of the cyber world, both from an attack perspective but also from a defence perspective. And I wear various different hats with different bodies around the world in terms of best use of AI, safe use ethical use and so on. And it's just a world that I love and I now find myself speaking about it an awful lot. I do two or three talks a week and I write books about the subject from different angles and viewpoints dependent on who the reader is. And yeah, it's just a world that I absolutely love.

And training courses as well.

Yes, indeed. 

Which a number of the team at Pro2col attended at your HQ.

Yeah, no, I do do a lot of training. We're a team of 130 people based in Surrey.

Lovely. And I'm very excited to have 40 minutes with you today to learn about your world and more about AI because I know, well, who doesn't know about the term AI these days? In terms of organisations that you work with, is it you're working with organisations around the world in every different sector. You're working with finance, retail, government, is it all of them?

So in terms of, so I should differentiate because my speaking is global. The business that we do is typically in the UK and it is across all sectors, but it's organisations that typically are data based or people process, you know, logging of any form of data and information are all typical companies and early adopters of that AI journey.

Similar to us being in the world of data transfer.

Indeed, yeah.

And we'll probably come on to that later on in terms of, we'll touch upon it.

I'm sure we will.

So from your perspective today, what is the current reality of AI adoption in businesses?

I think that most businesses are playing. I think that an awful lot of people have sort of stumbled into it by accident and not strategically, and that's a problem. I think AI is such a buzzword and organisations know that they need to address it. And I think one of the problems is you get somebody who's an enthusiast and they see a product and they bring it into the business and they go, look at this, we've got to use it. And then people waste time and money trying to fit their organisation into that tool. And that tool might be the wrong tool. And the other challenge is when it's driven by IT because this isn't an IT thing. It's a strategy and that strategy needs to be driven by the board and the board need to understand what's the business opportunity or what's the business challenge or pain that this solves. So I think it's a real mix you get other businesses that are trying to shut down use of AI because they're nervous, because they actually don't know how to do this safely and securely. So they go, this is our tool, we've got one tool and this is what we're going to use. There's a real, I was talking about this on stage last week and I was talking about the challenge of AI lobotomies and some people when you say AI lobotomy, think, oh yes, that's people getting dumb, isn't it, because they're using AI. And that's not it at all. And AI lobotomy happens within organisations where you've got really capable people who are playing with AI and they're using AI at home. And they know what it can do. And then they walk through the door when they go into work. And because they're boss has said, you cannot use these tools. That's when the AI lobotomy happened. And you've got lots of organisations with people who know that things could be done better and more efficiently and more cost effectively. And because of the policies, they're not able to. And I think there's a big risk there because I think organisations will lose really good, capable people. because actually the board hasn't taken a hold of this and moved it forward.

Is that because the board doesn't understand and they've made a blanket decision that this can't come into the business?

Yes, almost always.

Which isn't the right decision because AI is where we are going. So if you don't jump on it at all, then I'm guessing you're going to run into issues later on down the line.

It's just, the thing is if you're not doing it, your competitors are.

Yeah.

But from the outside world, you don't necessarily know which of your competitors are. your end user, your client doesn't necessarily know. But if your competitors are being way more productive and efficient, and I had it in my own business when we very first introduced a tool into our help desk and we were testing and we'd got some some people on the help desk using the AI tool and others not. And the people who were using the AI tool were able to do three times the number of support tickets every day. So you then roll that out across an entire organisation or think of that in other organisations and your competitor is able to do three times the amount of work with no less effort. I mean it's just, it's not sustainable.

No. When we talk about AI, I personally feel I don't know if it's the same or that you see. Where you have AI, a lot of people away from a business, they don't know something, they ask AI and they assume that's AI, they're going to ChatGPT, they're going to co-pilot to ask those questions. But from my understanding, that's a LLM. And often people just assume AI is LLM for the benefit for me and maybe some people listening, could you talk to the differences between AI, LLMs, machine learning, is it genetic AI? Agentic AI, yeah. And the differences between those.

Sure. So you're quite right. AI is the umbrella term. And artificial intelligence is generically systems that are created to mimic human intelligence. And that's not new. We've had AI. Well, in fact, Alan Turing started his work on AI in 1945.

Wow.

In 1950, we had the imitation game, which lots of people will have seen the movie about. And the imitation game was he posed, can I get a computer to fool a human into thinking that they're talking to a human? That was 1950, about 76 years ago. So generically, AI has been around all that time. There are different branches of AI and it's a big field. You mentioned LLMs. So an LLM is a large language model. So the likes of ChatGPT and Copilot and so on are LLMs. And the significance there is they've been designed to interact and respond to human language. Whatever that human language is, wherever you are, it understands dialects, it understands colloquialisms. So you're able to interact with it in the spoken format. And that's a massive enabler because all of a sudden people are able to just talk to tech. The generative AI will create new things out of the data that it's got access to. And then when we talk about agentic AI, an agent is something in AI that will do something for you. And the other buzzword of late is orchestration. An orchestration is you orchestrating agents. So you have a series of agents in your daily work life that does certain things and you tell them what to do and when to do it and what to do with the output. So you might have an agent, for example, that checks a competitor's price and then emails you. Or writes a sales report on a Friday evening at close of play or whatever. So all of these conflate to the generic descriptor of AI, which incidentally I like to talk to people about as assistive intelligence rather than artificial intelligence. And that's a big tip. If someone listening to this is thinking, I know we have to do AI, but we're just starting out on that journey, I think an awful lot of people go into businesses and talk about, we're going to introduce AI. And there will be a proportion of your employees who are really anti that and are really nervous about that. And they clam up because they think, oh my, you know, I'm going to lose my job or whatever. And my advice is don't talk about it as AI, talk about it as assistive intelligence. And if you change the conversation, if before you even say to your people, we're going to introduce AI, if you have a conversation that says, if there were no budget constraints and you could have an assistant to help you in your job, if you could give the stuff that bores you, and is repetitive to somebody else so that you can be more human and spend more time with your customers, what would that assistant do? And people just have so many ideas because there's all of those things in everybody's job. And guess what? The answer is AI. But you get a totally different response by asking that question rather than fronting it with we're bringing AI.

And even just chaining it to assistive intelligence, that word assistive completely changes the meaning of it.

Yeah.

I'm obviously been on Rob's course, so I'm thinking here, sitting here thinking how relatable that is to the automation that we have in the platforms that we've got and the agents that can sit there and they can do certain processes. Obviously, the Agentic AI has levels of comprehension and ability way beyond the agents that we've got that move data around, but fundamentally it's not dissimilar in terms of maybe the evolution of the technology. We're not quite as cutting edge, should we say, as some of the AI capabilities. But yeah, there's that juxtaposition between what's AI and what's automation isn't there. So, and you know, we're sort of sitting in a similar space, I guess.

Yeah.

Who are you finding is the driving force behind? the AI adoption across enterprises? Is it the people set at the top, the board, the SLT team, or is it the teams below? I guess actually from what you're saying, it's probably the people in charge.

So I think there's two ways to answer that question. Because I think the ones who are doing it successfully, it's the board, it's top down. You know, it's I think if I map out AI maturity, then there's a couple of things. One, who on the board, who on your board has responsibility for AI? And it's often not the same person who has responsibility for IT or tech. So is there somebody on board level who's got responsibility? The other real telling question is what's your AI budget? And so many people say, well, it's an IT cost. It's not. It's not. I think treating it separately is really important and is a tell of, as I say, the maturity of a business on that AI journey. And that's because the board have driven it. They've taken ownership. They've bought people in to help them on the journey.

So does it sit inside like an operational budget or maybe an R&D budget. Where would people usually?

I'm sorry. I mean, we have AI as a totally separate line.

OK. You've got a decent understanding of what those costs are going to be. If somebody's starting out on their journey, I guess.

And it's massively changing and it's changing quite fast. The, you know, the common models that most people are playing with all of a sudden are going to start to get quite expensive.

Yep.

I did an article about this recently where I called it the Sky TV-isation of AI. And what I mean by that is when Sky TV first came out, you paid 25 pounds and you got everything. It was just like, wow, this is amazing. And now they reckon the average monthly bill for Sky TV is about £140 a month.

Yeah. The many conversations I've had with my mother about reducing that bill.

But I think that's the journey. And there were so many, there were so many, I mean, I use so many different AI tools, but that's not, you can't give all of those AI tools to everybody because everyone is, another $10 or $20 a month. So you then need to start thinking about, right, what is your AI budget? Even just in terms of tools that you give people on their desktop, what are you budgeting? Because it's very easy to go, I need Copilot to do this and I need Claude to do that. And then over here I need Granola or Whisper or whatever it happens to be. And all of a sudden that £60, £70 per user per month. But then we get into that whole tokens or credits. So Copilot, Copilot released Cowork. So Cowork is actually came from Claude. And this is that orchestration of agents piece, which you can now do in Microsoft Copilot. When it first came out, you had a cost as of the 1st of July. It's based on tokens. Yeah. And if a if a 17 slide PowerPoint costs $5 to create. And then you start to roll that out across users. And don't forget, this is on top of your standard licensing. So you've still got to license it to have user licenses. So I think understanding of cost of all of this is really important.

They've invested an awful lot of money. They need to start recouping some of that investment, don't they? And it begs the question whether AI is going to become too expensive. And certainly if you don't have that strategic input in terms of making sure that you're selecting the right tools for the right projects, for the right problems that you're looking to fix, then yeah, the spread and the creep and the costs could become a real problem for organisations.

I agree, but I think it's also part of that maturity in terms of you've got lots of people who are just using the tools that they can just subscribe to themselves and all of a sudden are in this place. Whereas actually you need to be working with people who understand this to go, okay, for example, one of the AI tasks that we do, which is a relatively heavy demand, but doesn't need the intelligence, we've put on a llama server on site.

Okay.

So llama is just another AI that we've got running locally that isn't costing us loads and loads of tokens. But you need people to be able to come in and actually work out what it is that you want to do and what are the pains and how do we do this efficiently all the time considering that safety and security and ethics and cyber risk and all the rest of it, that's really important.

Some of our biggest customers are very much a top-down approach to wanting to know what AI is going to be available in or around or associated with the tools that we're specialising in. So it's an area that we focus quite a lot of effort and time into. And I know that we have another podcast. specifically diving into the details of how these tools can either be leveraged around the periphery of it or some of our vendors are starting to build out some capabilities in that space.

Yeah. There's a lot of talk at the moment actually with all the vendors in the MFT space, like you say, are all doing something bubbling away behind the scenes, but they are all. AI is on their roadmaps and I think that things are coming. I don't know what.

The easy wins are the reporting, automating reports and getting deep into that and aiding compliance in part through the reporting capabilities. But I've seen some interesting stuff from various vendors around automating some of the workflow tasks or helping with the creation of the workflows. That'll be interesting to see how that evolves over time.

I think when you so big organisations in the industry like Palantir, for example, they coined a phrase called forward deployed engineering. Which forward deployed engineering basically means an AI specialist goes into a business and they walk the floor and they understand processes and they map the steps of the process and they understand where are the pains and you need to do that before you work out what it is that you're going to actually automate or put into AI.

Interesting. Just if I can in there, it sounds like the just-in-time principle is going back 40 or 50 years ago, but in a modern context.

Yes, it is. And in fact, I was reading a fantastic book just recently called The Algorithm. And the algorithm was written by the president of Tesla. And it's what he learned from Elon Musk and has now taken to all the businesses that, so he left Tesla and he was at SpaceX and he went to General Motors and Lululemon. But Elon Musk has this business model. And interestingly, one of the fundamental points is about process mapping and understanding the process before you do anything. And his approach, which I love, He goes and gets everybody in the team to write post-it notes of every single step and then stick the post-it notes along the wall. And then his challenge is, okay, circle in red what the customer's paying for. Because everything else is for you or we've just got fat and we've got lazy and our systems don't work properly. And then you look at fixing and simplifying some of those things before you automate. And then I think the evolution from that, where a lot of the AI work that we're doing with customers is building an MCP layer. And what we're then able to do is link together the different tools that are in use in an organisation. So if I take it, you know, when we did this for ourselves, we've got our help desk system, we've got our account system, we've got our CRM system, we've got voicemails coming in. And what the MCP layer does is it surfaces information out of all of those to whatever somebody's AI companion of choice is. So if you use ChatGPT, you can use ChatGPT and you can go, I've got a meeting with ramsac. give me a quick update and it will talk to all of those systems and answer you. might be chatting to it or you might be chatting to it in co-pilot or whatever. Sure. But that linking into all of the systems so that you can hold it. Now I think that's a big leap forward because what that also removes for a lot of people is that needs to replace systems. They've got systems that don't quite do what they want them to do. And they're looking at spending lots of money. And actually, AI is a great way that we can just integrate into that and talk to that.

Interesting way, yeah, consolidating all the data so you can query it in the ways that you want to. So the MCP is running internal to you or you've got that internal.

Yeah, so within our tenant. Yeah.

Can I ask a question? What does MCP stand for?

So it's model context protocol. But it's a standard term for a technology in AI. I probably should have mentioned it when we were going through the different definitions previously.

I was like, should I know this term? But I'm glad I checked because I did not know the term.

And most people won't. And most people will talk about. Yeah. But it's, but people talk about all of these terms in AI without actually going, okay, what is, what is it? What does it do? How is it going to help me? Yeah. What is GPT in ChatGPT?

I actually don't know.

So GPT is a generative pre-trained transformer.

Well, there you go.

I didn't know that. Something new for today.

Just touching on sort of AI systems and how businesses are adopting AI tools. You also mentioned earlier, there are a lot of AI tools out there to choose from. When you're working with businesses, do you tend to recommend using multiple AI platforms depending on what they need to do? Or can you see a world where there's going to be one central AI system? I'm thinking in terms of businesses that we work with, they may have two or three different file transfer systems or maybe multiple systems in place. Do they then need an AI system per system or could there be a centralised?

So and I think this is really interesting because there's a couple of ways to answer it. On the one hand, everybody wants to know what's the AI that we should use because people would love just to go, okay, we just need to give everybody a license for whatever the tool is. And I don't think that's realistic because I think different AIs are more capable some AIs are more capable at doing some things than others. So your coders will all want Claude, probably. But your creatives aren't going to want Claude. Because although Claude's now got the design function in it. It's not creative in terms of image or video generation or whatever. It just doesn't do that. It doesn't go there. So your marketing teams and your creative teams are likely to have something else. And I think it's working out what is the most appropriate tool. Brings us back really to that question about what's your AI budget. Because if you had a licensed budget as a cost per person, it actually makes it much easier to work out ok, we're not going to have this license drift in spend. We're going to agree that we're going to spend however many pounds of dollars it is a month. If you've got a finite budget, then it becomes much easier to work out what's the tool that person really needs. The other part of the answer is back to the MCP piece, where I think what that does, and I think every organisation will end up with this MCP layer. Some people are calling them AI applets. So it's an AI applet that basically sits in your organisation that talks to all of the tools that you happen to use in your organisation to make it seamless. And I think we will see much more of that.

Yeah, interesting. I could have answered that question for you. Not often I can answer a question, but having been on Rob's training, I've seen a presentation and there's this wheel, maybe a very small advantage. I remember seeing this wheel and there must have been, I don't know, guessing 50 or 100 different AI tools. And having been in Guildford and came back home, I spent a few evenings researching different tools, some quite interesting sales tools, BDR type stuff. which is quite it's quite interesting, but we haven't I got I got lost in some various other rabbit hole and haven't quite resurfaced on that one yet, but.

And again, it's working out. So again, we've moved our whole CRM into an AI platform, but we use more than one platform. So we've got a CRM system which is based in Attio. We integrate to it through our MCP so I can literally just chat and I can talk to it and I can say just had a phone call with the client and we've agreed that we're going to chat again on Friday and the CRM system just updates. But to your point where I was saying when you came in to do the training session, there were there were hundreds of AI apps. And my advice is always the the advice that AI will give you is not bias. There is bias in AI, the wrong sort of bias of the race and sex bias exists. And that that's terrible. And we're trying to to work on that. But in terms of technical advice, You ask any AI, so you say to Claude, what is the best AI to do video creation? And it might say, use Grok or use Nano Banana or whatever it happens to be. So the advice that it gives, or even, this is the tool that I've got, this is the application that we use, this is my industry, and it's not doing what I want it to do. Is there an alternative? Is there something that we do it better? It's a really good starting point.

Yeah, some very specialist tools out there. are. A lot of our clients that we talk to are concerned about guardrails and the risks that come with adopting AI within a business. Do you hear that a lot from people that you talk to? And so what's your advice to how they leverage those risks?

So I think, yeah, I mean, it's really important, really important. And I think one of the issues is it's so easy to start to play with this. When I look at Copilot as an example, when Microsoft were first working on Copilot, they were going to make it a minimum of 200 users. And that was actually great because when you look at the licensing cost of 200 users, that's a reasonable amount of spend. And if someone's spending that amount of money on IT, they treat it as a project. And when you do a project, you do a scope and you do training and all of that stuff. And then they changed their mind and went, actually, no, you can just buy a single license, which is £26. What then happened is you got people who bought licenses and just turned this stuff on. without doing the groundwork. And I think of it, there's a whole load of data preparation and security and labelling and permissions that needs to be done. And visually, I think of it as preparing the soil before you start to plant a crop. And yet people just chuck seeds into this unprepared field, hope for the best, have all sorts of problems. So the preparation piece is really important. And I think one of the things that will evolve, I mean, the AI part of my business has grown massively. And one of the things I'm seeing is, so we started life as an MSP. We became an MSSP on the cyber side of things. And I think one of the things that we will see is that organisations will start to turn to whoever it is that looks after their security and their IT systems to do AI. Because to the point that you raised, yes, there's concern. There's also an awful lot of one man bands or two man bands who have been early adopters of AI and are really quite wizzy and there's a fear because they're going out pitching ideas to businesses and they do the work but then they move on and there's only them and they're doing another client and actually you then have a data problem and who's there to sort it out or to these systems update and change so frequently that there's always an ongoing maintenance because a new model comes out and all of a sudden something that was working last week isn't working now. And so I think to your question, yes, people need to be cautious.

Yeah.

And I think that's right. And actually, I'm delighted that people are what what worries me more is the people who would, as I say, just chucking seeds into yeah, into the unprepared field. Metaphorically, those are the people who worry me.

Yeah, that's the nature of what we do and being involved with data. A lot of our clients are wary of introducing AI.

Maybe, some of them, maybe not all of them. I mean, they're definitely leaning on us. So we had a customer event last year and it was extremely apparent at that event, which you were at.

Yeah.

Their #1 question at that time, and there were some very senior IT executives in the room. And because where we sit in an organisation, they were very much in the IT function of the business. They were banging the drum, what's happening? What are you doing? How can you help us? we've got mandates from the board, we need to be introducing AIs. And to the extent some of those customers were saying, if it doesn't include AI, then it's not as important as it probably should be for our business. So you talk about being the trusted advisor, I guess, for your customers, that's sort of fundamentally where we're at and why we're putting as much time into understanding capabilities around AI, which is why we came to spend some time with you, but also some esteemed colleagues, Richard and Dave, putting some time into understanding as much as they can around AI and what that means within our space so that we can then help our customers to understand and be there to take them on that journey.

I think the other interesting piece in that is customers want to know that you are doing AI. It's like, how are you harnessing it? How are you, how is the product of the service evolving? And there's a theme in the industry this year about transparency, which is just that. You know, let's be transparent about what we are doing with AI and what we're not. I think the next challenge, and this is going to evolve, is how does that ultimately get charged for? Because the more that you do with AI in terms of a product or service, that's going to start to cost.

Agreed.

And people's. So another question I sort of alluded earlier to the AI maturity journey. And one of those other steps in that journey is what is your post AI business operating model? And what I mean by that is talk to any business and say, is AI going to change your business? Is it going to change what you do, how you deliver it, what your customers expect? And virtually everybody says yes. When you then say, okay, have you sat down and redesigned what your organisation's going to look like in response to that, very few people have.

I'm sure. Yeah, it doesn't surprise me. I mean, we've got software vendors at the price point and the budget point is an interesting one. Vendors are looking at building in capabilities into their platforms, but then it's the, how much is this going to cost from a tokenisation perspective? It might be charging for a license for it. We can quantify that relatively easily, but as soon as lots of throughput starts coming through the AI model that it built into it, how does that work? as that being paid for, which commercial route is that taking? It presents a number of challenges for us as a supplier into these organisations at the very least.

Absolutely. And you look at that as well and you think there are people who have reacted because they need to get AI into their product, but they might have gone the wrong way. And now they're going to be hit with this massive cost. And all of a sudden, you're paying to use their service and you're going to get a monthly token bill on top of your license to use the system. And where does, you know?

And then from a budget point of view, you've got to weigh that up, haven't you? You've got your license and token versus person. Where do you go? So yeah, it's going to be an interesting next 12, 18 months to see how this all shakes out.

How easy, can you measure return on investment with AI? I know, well, James mentioned we speak to clients and some of them are targeted on AI, they're building strategic proposals to integrate AI into their organisation. How or in your opinion, how could they go about proving return on investment on this?

I think on the ROI question, it depends what you measure. and different organisations will choose different measures. some people, and I think this is wrong actually, but some people will do a literal headcount calculation. It's like, well, we've bought in this AI and therefore it saved us X number of people. My view is that that's the wrong model because for a lot of people that's just not the way it will pan out. it should, for me, when we look at what AI has done within our business, we've freed people to then do other roles. And, if we measure, so one of the things, we had a role on our help desk, for example, and it was two people. And what they were doing, when people log support tickets, They either e-mail them, they either put them on the client portal, or they phone up and they leave a voicemail. Tickets could be raised in any of those three ways, or just phone and speak to somebody. So the first three things that I talked about, we'd got two people who every day were going through the incoming emails, they were going through the tickets that had been raised on the portal, They were listening to the voicemails to then log a ticket to allocate it to the right person in the right team to deal with it. Well, AI does that for us now. You know, it listens to, and by the way, of those voice calls, most of them, or quite a large proportion of them, were sort of pocket dials. But you still have to listen to the message to work out, is it actually someone, oh, sorry, Yes, I want help with this. And AI does all of that triaging for us now and then passes the ticket to a human to actually do the support call. But those two people that role was replaced are now able to actually do the support. And nobody in my support team, as an example, joined my organisation to do triaging. They did it because they enjoy tech and they want to answer questions and and so on. So I think if you look at what's the problem you're trying to solve, is this about efficiency? Is this about capacity? Is this, what is the AI looking to do? And I think if you start to focus on that and then measure that as an ROI, that's much easier to start to calculate. It's like this process that we've introduced into our systems now does all of this in X amount of time. We can calculate and we can report on that.

There's a few other angles as well. I think the speed to response, if we think about from a marketing aspect, I know that we track the speed at which we respond to a customer. or the prospect at that particular moment. And the resulting likelihood of them becoming a customer is, the correlation is very, very clear to see. So there's the speed element of it, but there's also, I mean, you talk about the support people. So you take them off the basic support element, but potentially they're in a billing role at that point and you're charging them, charging customers for their time. So there's another aspect to it. It's not just necessarily cost savings, but the revenue generation as well. And the other thing is your support people are happier, theoretically speaking. They're going to stick around longer. So there's all these incalculable, non-numeric sort of things there that People stick around in organisation, although we're lucky enough to be in a have a really good retention rate on our staff, and it just makes a huge difference for the customer service element. So that's, yeah, that's it. You've opened a can of worms there. Apologies.

But I think, yes, absolutely. So I guess the point is look at the business outcome and measure the business outcome as the ROI.

And as we look ahead. Five years, because I imagine a lot's going to probably change in five years, so I won't ask you to go 10, 15 years, 20 years. Where do you see sort of AI shifting to? Do you see any dramatic changes?

I think that in essence, right now AI is a tool that you open. Whatever AI you're doing, at some point it's a tool that you open. And it will just become embedded in the things that we do. So you won't open it. will just, it will be there. It will be a process. Already, as I say, I can pick up my phone and I can speak to my CRM database. I've got a number of AI tools in the background. I've got my meeting recorder that records my meeting and then just automatically passes that into the CRM, creates tasks. That's an example of the early stages of this integration where it will just be part of. Daily life. Yeah, it's core infrastructure that you can't. I use another app called Whisper. And Whisper lets me just talk to whatever device I'm talking to or whisper to it, which is where the name comes from. And It went down yesterday. I was using it and it went down. And the panic, because all of a sudden, the beauty of it is it's so much faster than typing. And I've got four screens on my desktop and I'm running various different programs and I just click on one and I just talk to it and I tell it what I want it to do. And then I click on another one and I tell that. And that ability to multitask is immense. but really noticeable when it stops working. But we are in these earlier stages of some of this stuff, but I think it's indicative of the fact that AI will just be integrated into our flows and our process.

I need some whisper in my life.

Yeah, it is amazing.

I can see why it's amazing. I mean, even on like WhatsApp now. A lot of people just talk and send the message because no one wants to type anymore.

And the difference with this, so you have got that sort of ability to dictate, but what this does, so this is an AI dictator, so it understands you and it learns your shortcuts or your mispronunciations or whatever it happens to be and just auto-correct. And you say to it, this is what my style is. So are you, old and do punctuation and capitalisation in text messages, or are you young and do it all in lowercase? That's not a setting, but it is an example.

Can relate to that. I'm fascinated with where this is all going. And it reminds me that I do need to spend some time looking back into some of the tools that Rob had suggested might be a starting point. I'm going to go back to that wheel and have a look at them, see where Whisper is on there.

It probably wasn't on now actually. No. That's another thing in terms of how AI is just evolving.

That was not much more than six months ago indeed. Yeah.

As a final question for those listening who want to cut through the noise and with AI, where should they start? What's one thing today that they could go away and start doing to adopt AI effectively within their business?

I think firstly, make sure this is a bored conversation and let's move away from AI just being abstract and focus on what's the one thing. If we could solve this one process, this one pain, one problem, what would that be? And start there and then take some advice talk to specialists like Ramsack who can actually go, okay, well this is the journey to go on. And to build on that, even that post-it note exercise, which I just think is fantastic and anybody could do in any team is a really interesting way because when you start to do it and you map it out, you go, really, we do it like that? And so many organisations of all sizes processes have just evolved over time. And I think the worst thing that many people do is just automate all process. So work out what is it? What could we do better? And focus on that and have a I work with some clients where they have a problem 1/4 and they go, okay, for this quarter, this is the problem that we're going to aim on resolving. And I think that's quite a nice approach as well, which doesn't overwhelm people.

Yeah.

Great idea then.

I have one more question before you leave.

Before you do.

I can't let you.

Before you do. It was an opportunity for a shameless plug. That last question was an opportunity for a shameless plug. So to camera, you need to go and check out Rob's books. I've read a couple of them. international bestseller. There are some great ones on there. AI strategy being one of them. It's a good place to start.

If you've got some links, we can leave them in the show notes. They are easily accessible.

Yeah. Sorry.

That's okay. No, I just, I can't let you leave without asking what's the coolest thing you've seen or done with AI?

Well, okay, so here's another shameless plug. So my website, my speaker author website. It's called thoughtprovoked.co.uk. And I've just completely rewritten that entirely in AI. So go look at thoughtprovoked.co.uk. The entire project is AI. A lot of it, me speaking to it. is incredible.

So you still got all the logins to get into all the data in the back end for past customers.

Yeah. So when you go to it, It will say, you're a past customer. I recognise you need to choose a new password. And then you get into the resources that you got from coming to one of the training sessions. All of that's ported across. Okay. But again, all through AI. It's astonishing. I'm not a code, you know, I'm not a web coder. But I think everybody is becoming a coder is the reality. You know, everybody in my business, whatever their role is, is now using AI to solve problems.

Well, that is all my questions. And I just want to say thank you again for joining us and giving us the time today to share your world of AI. It's been fascinating.

Pleasure.

We hope you've had fun.

Yeah, it's been great. Really enjoyed the conversation.

Good. And before you go, where can listeners find you?

So either, so corporate website is ramsac.com, R-A-M-S-A-C. Or my speaker, author, website is thoughtprovoked.co.uk, my latest AI project.

Thank you so much for listening. We hope you enjoyed the conversation as much as we did. You will find Rob's details in the show notes below. And if anything from today's episode sparked a question or got you thinking, feel free to reach out to our team. Our contact details are below also, and we are always happy to have a chat. Make sure you hit subscribe so you don't miss out on future ones like this one. And I can tell you, in the next episode, we will be continuing the AI conversation, but bringing it much closer to home, looking specifically at AI in MFT. Thanks again for listening and we will see you next time.