The Social Media Takeaway - Louise McDonnell

Work Smarter with AI: Increasing Productivity and Saving Time

• Louise McDonnell • Season 1 • Episode 42

Share episode

In this week’s episode of The Social Media Takeaway, I’m joined by Alastair McDermott, a consultant, author, and podcast host with over 17 years in business and tech. We dive into the practical capabilities of AI in everyday work, from handling repetitive tasks to enhancing productivity. Alastair shares tips on getting better results with AI, like his “GOAL” framework for prompts, and we explore how AI is shaping our professional lives. If you’re curious about how to integrate AI into your daily tasks effectively, this episode has some great insights.

Get Alastair's AI prompts collection - including his productivity shortcuts, podcast workflows, and writing formulas at aipowered.biz/louise

Get your AI prompt toolkit here! 

More about Alastair: 

Website
Linkedin
AI Frameworks

If you enjoyed this episode, be sure to subscribe to my podcast because more like this is on the way!

If you'd like to book a call to see how I can support you head over to my website here. www.sellonsocialmedia.academy/hello

And please connect on social media and let me know what you thought of this episode!

LinkedIn
Instagram
Facebook
Facebook Group

Check out my 2024 Social Media Content Planner & Guide on Amazon (Amazon UK) (Amazon USA)

Louise McDonnell:

Welcome to the Social Media Takeaway. I'm your host, Louise McDonnell, and this is the show for business people that want to know how to use social media. more effectively to drive sales and leads for their business. If you enjoy the show, please make sure you subscribe. And if you know anyone who may benefit from this show, please make sure that you tell them about it. Be also sure to find us on social media, so you can maybe give us some feedback, ask some questions. And in general, connect. I would love to connect with you. You'll find me under@sellonsocialm, on Facebook, Instagram, and TikTok, And I'm more likely to be hanging out on LinkedIn, under Louise McDonnell. So today I'm delighted to be joined by Alastair McDermott. Alastair is a consultant, author, and a podcast host with over 17 years experience running a business and consulting for other businesses on technical and business challenges. Alastair brings a unique blend of a technical experience, and business He has a background in software development and an understanding of both technology and business challenges. And he speaks regularly about business strategy, marketing, and AI. And I know in the last while or so, he has really moved into focusing solely on AI. And in fact, he's just brought out a new book, which is called An Absolute Beginner's Guide to Using AI. Alastair, you're very welcome to the show.

Alastair McDermott:

Hi Louise. Thank you for having me on. It's great to be here.

Louise McDonnell:

It's great to have you. In fact, you're probably the reason that I started podcasting because you're the one who said it to me probably about this time last year, maybe a little bit after this time last year.

Alastair McDermott:

Yeah, absolutely. And, for anybody who doesn't know, I'm big into podcasting. I have one, podcast that has 175 episodes and I have probably five or six more podcasts that have less, but I've been doing podcasting for a while. And I really love it. I think, I think it's a great format. I'm delighted that you started this.

Louise McDonnell:

Fantastic. And actually, now that you've mentioned your podcast, how can people find it?

Alastair McDermott:

They can look for, my newest podcast is called the AI powered business leader. So if you look for that in your podcast app, you should find that, or you can visit aipowered. biz slash podcast.

Louise McDonnell:

Very good. Very good. So tell us about this journey to AI. What has brought this about Alastair? Tell us

Alastair McDermott:

what you're thinking. Well so many moons ago I was in college doing software engineering and I did a course on AI. And the AI that I worked on back in, you know, 1999, 2000 is nothing like the AI that we're working on now. It has almost no relevance to, to where it's at now. But I was doing software in college, so I've always been that kind of techie engineering type person. And then I started my own business maybe 17, 18 years ago. And I've been in the web design and online marketing world for a long, long time, like yourself. And so recently about. I'd say 18 months ago, we're coming up on the two year anniversary now of chat GPT 3. 5 being released, which was kind of the, the landmark moment in, AI, suddenly people putting their heads up and taking notice of, Hey, this is actually getting quite good now. So about 18 months ago, which is about six months after that. I was talking to a friend of mine and I said, look, I'm leaning more and more into this AI stuff because I think it's going to be a big deal. And he said to me, well, like, are you doing, are you becoming an AI person or are you going to keep doing the marketing and kind of do a little bit of AI as an aside. And I said to him, I think that if I continue to just focus on web design and marketing. it would be a little bit like if I was sitting on the beach drinking a cocktail while there's an oncoming tidal wave, and I'm just purely ignoring it because I think that AI is gonna be that big and have that much of an impact, particularly on the world that I'm in, that I think that I need to focus on it. I need to, okay. What you need to do then is you need to stop doing everything else and just do ai, which is why I stopped my podcast called The Recognized Authority after 175 episodes, and I love that podcast so much, but I needed to focus on the AI stuff So yeah, so I'm all in on AI. I think it's, it's not there yet. Like, I think it's really good, but it's not great, great. And I think that it's going to improve rapidly. I think it's something that we should figure out how to use, otherwise we'll get left on on the wayside.

Louise McDonnell:

And actually I actually didn't realize how much I use. I'm only really into, I think like the chat GPTs, like I'm not really into image creation. We use a lot of, video editing AI tools, which are amazing. But I didn't realize actually already how much it's actually ingrained in everything that I do and even in my coaching courses, someone said. Are you going to use AI? And I'm like, I already use AI in every single step. But I didn't realize that that's kind of the USP that I have. So, yeah, I think it's just amazing the way it's just become and parcel of what we do. So tell us, what's your take on it, Alastair? Where are we and where is it going to go?

Alastair McDermott:

Well, we're right at the start of something that is incredibly powerful, but still incredibly limited. Like, for example, Anthropic, who make the system called Clod. Clod. They released last week a system called computer use, which is where Claude can control this little virtual computer. And I tested it last week, put out a video on LinkedIn, if anybody's interested in checking that out. And it's really clunky and it doesn't work very well. It keeps running into problems, but you're sitting back, you tell it to do something and give it some instructions. You sit back, take your hands off the keyboard and it's doing it. No, it's not doing it very well. It's not doing it very fast. And it runs into problems like the cookie pop up comes up and it's not able to get past it. But if you say, Hey, you know, click okay on the cookie pop up and then you'll be able to do it and you tell it, you teach it, it's able to do things for you. And so what that means is that at some point I'm going to be able to tell the AI, Hey, take over either my desktop computer or maybe like a little virtual computer that's running on top of my desktop computer. And I'll be able to tell it to do something. And go away and have lunch or have a coffee and then come back and it'll be done. So I'll be able to say, Hey, go into Riverside and download the latest episode of my podcast and get it ready for editing, or in fact, can you do the editing for me and then upload it to my podcast host and take the transcript and turn it into show notes and publish that to the website as soon as it goes live on the podcast host, and we'll be able to do that and it'll be able to do it for us. And that is going to change a lot of things when it's able to do that without making mistakes. Right now it's a bit like the self driving car. If anybody's been in a self driving car and put on the self driving mode what happens is it drives along and then it says, Oh, I need you to take control. So basically an intervention. And right now the computer use is like that. It needs you to have an intervention to fix things. But at some point, it's going to get to the stage where it doesn't need the human to have as many interventions than it's going to need. No intervention is just going to be able to work for us. And then we're going to be able to ask our computer to just do some work for us while we go and do something more valuable. Maybe we can go and have an in person meeting because we've got time for that now. So we're not spending time doing bookkeeping or whatever the really boring task is that we don't really want to do.

Louise McDonnell:

I can see it working for the boring tasks where you don't need human input. maybe I just find like, this is one of the downsides I think of using, let's say, ChatGPT, is that sometimes I think people go to it and they say, write me a five email sequence to encourage people who have signed up to my webinar to turn up live.

Alastair McDermott:

And I think if you do that with, with ChatGPT, certainly. It'll give you like a whole lot of rubbish, like, so my way, can do, but let's say you say, and by the way, here's the best practices for writing a five email sequence to get somebody to turn up. And by the way, here is the transcript of my last webinar. So you know what it's about. And so take the best practices for writing that email sequence and take everything that you know about my webinar. Now write me an email sequence that's much more customized to my content and follows best practices. Suddenly you'll find that that's actually pretty good.

Louise McDonnell:

Yeah. Do you know what Alastair? I think that my experience of using it is that even when you plug in all that, you'll get a much better response If you sit down and you go, let's come up with the best subject line, let's come up with the best this, let's go. And then like, I have conversations with them, write something for me. No, no, I don't like that. I think I need to explore more of this. And sometimes it goes back to, yes, you're right. Or no, I don't agree. So I think, I think it's really powerful, but I think trying to get it to do too much at the one time. So if you say, write me a landing page, it'll give you a rubbish, not rubbish, not rubbish, but not good enough. Whereas if you just focus on the first section, second section, the third section, then I think you can come up with something that's a lot more, a lot more valuable. So tell me the difference, what's the difference between ChattyBT and Cloud?

Alastair McDermott:

Well, they're entirely different models and so they have different personalities. I've got different writing styles. I've even noticed like the different versions of Claude from one to another, the writing style, the English it gives us back, the written word is very different. And in fact, I used to love the way that Claude wrote a couple of versions ago. And I didn't like the last version as much. This latest version that we've just got recently. I haven't used it enough, but I think it's a lot better. It's a lot closer to how I personally would write. But for all of these, like I have very specific writing instructions that I use. So read it out to you just so people get an idea of what I say. So this is the specific instruction I give ChatGPT. I say, don't forget to write like me. I write in a direct, practical, and actionable style with no filler or fluff, and without adjectives. I write in UK English with no jargon, in simple English at a 12th grade level, in a conversational manner that should convey a sense of openness from the author. I mix the length of my sentences, I write from the first person perspective, I address the reader directly in the second person, you, and include we to establish rapport. And I often say this is important because to explain the value. To the reader. And I never, ever used the word delve, demystify, foster, leverage, or utilize. I use examine, encourage, use instead. I always use those and I never say, remember before making a point. I just say it. Now that is literally what I have in my instructions for right? Like me. Okay. I had to read that out cause I wanted people to get, okay. Like, there's a lot of instruction there. but when you give it some really specific instructions like that, it's able to follow them. And that helps a lot. And it's still not perfect. And you know, every so often, like, it's really annoying for us who don't speak American English because we have to remind it about the spelling because it will forget over time, you know. It goes back to the American spelling. And you know, it really loves to demystify it.

Louise McDonnell:

I know, I just love, I'm laughing so much. Please do not use, I have the, I think I said, please never, ever, Ever, in this version of ChatGPT, my paid account, ever used the word, what do I hate the word mastery and unlock.

Alastair McDermott:

Yeah. And, the other problem is that it's a bit like saying, I want you to make sure that you do not think of a pink elephant. And of course, it's impossible not to do that. So it's like when we give it instructions on what not to do It's very hard for it to follow a negative. And so it, it forgets it. It, you know, like these are imperfect systems by definition, but they're still incredibly powerful. And it depends on the way that you use them. And, and you really do have to look at the output. One of the ways that I when I'm talking to people about, you know, getting started, the very basics. I give them like an acronym for writing prompts and, it's goal, G O A L. And the G for goal is what's your goal? What do you want to, what do you want to say? The O is output. So what's the output format? So like sometimes you can tell I wanted the form of a table or I wanted an Excel spreadsheet, like CSV or something I can import, or I want to that as a HTML code or whatever it is that you want to get and then A is any additional context. So have you got any other information you could have, give it that will help it to give you better output. And then L is actually look at the output. And this is the one I see people make a mistake with, because you have to look at what it gives you and then make changes if it's not good enough. What so many people do is they just blindly take what it gives you. And that's the real problem. that's when you, you see the stuff that really sucks getting posted on LinkedIn or wherever else is because people are not looking and actually sanity checking. it can make mistakes. In fact, one of the tools I use a lot. Is is a coding tool and it compares to text files. So it compares like you can put two text files left and right side in, and it just highlights the changes. And I use that sometimes when I put text in and I ask it to fix something, I use it to make sure it didn't change anything else on me, because sometimes it does. Sometimes if you mentioned you know, like summarize or something earlier on in the conversation, suddenly you find there's a whole chunk of your document missing that you didn't realize. And it just disappeared because you had mentioned summarize. So there's things like that, that, you know, you need to get to know how the systems work and why they work in these ways. And nobody, by the way, truly knows how and why they work the way they do. Even the people who make them like they've, they've talked about things like, we don't really understand why, when you tell it please take a deep breath and then work through this step by step. Why does that actually make the output better? Well, we don't really know, but it does make the output better. that's been shown in, in so like, there's, there's strange things like that, that, you know, these neural networks and things, the technology behind it. not everybody understands, not anybody understands some of it. So yeah, but the more you experiment and use different systems and the more that you play around and spend time. Learning the more value you'll get from it. And then you can get to the point where it can do things incredibly quickly and incredibly fast. And that's what I like about using these things is it lets me become much more productive and create things much quicker. So I'm able to get a lot more done and create content and create good content, not just kind of crappy content, but actually creating good content based on the stuff that I'm able to feed into it.

Louise McDonnell:

Exactly, you know, and actually, my chat to BT is, is definitely learning, I put something into it yesterday and I asked it just to check something for me, you know, say like checks and checks from grammar. I don't even know if I had written it, but I was getting it checked. And it came back and he said, I've just changed the word. I think it was mastery. Cause we know you don't like, you know, you don't like that word. And I said, I guess you're right. I just find it's learning all the time. And I find it's getting more like my style because it's getting more and more used to it, which, and as you said, like, It just makes you way more productive and I think as well that like at five o'clock in the evening when maybe you'd be too tired to do a task before, now you know you can do it because it's not going to take that long and the quality is still going to be pretty good. Okay, practical tips for applying AI everyday tasks. So you talked about that prompt, which I absolutely love. Any other tips there, any other prompts that you'd like to use?

Alastair McDermott:

Oh, well, I, I've got so many, I'll just pull up my spreadsheet then. Okay. What I can do is I can share with you, I'll, I'll make a page on my website. I'm going to put it at aipowered. biz slash Louise. Okay. It's not there while we're recording this, but it'll be there before anybody listens to this. And you'll be able to download my prompts. Okay. Any prompts that I talk about here. some of the ones that I use are like the ones like the write, like me, and I use that write like me a lot. And I also tweak the grade level. You might've heard me mention in simple English out of 13th level grade level or 12th grade level. And I changed that between 8th and 13th, usually 8th and 14th. I actually don't really know, like, cause we're, we're not American. I don't know. Like what's 12th grade? Like what age is somebody in 12th grade? I have no idea, but I know that that will keep it to a, like, that will simplify it we'll make it not to jargony, it kind of reduces the the overall word count. Just in terms of complexity. So, and it's, it's much better. Like it's simpler. It'll do. What we found is that when people read, even, even very smart people reading English, that's written in simpler language, they're able to comprehend it better and faster because everybody is scan reading everything online. So writing more simply is good. I have some that would be useful for somebody who does something like podcasting. So I have a podcast prep conversation with every podcast guest who comes on, I have a conversation with the beforehand, Hey, here's what we could talk about. And it's usually the first time I'm meeting them as well. So I have some small talk with them at the start where, you know, you just get to know somebody. So what I ask the AI to do is I say something like, I'm gonna give you the transcript of my introductory call with the potential podcast guest for my podcast. I want you to skip the initial small talk, identify the podcast discussion, summarize the discussion topics, determine what angle we're gonna take on the overarching topic suggest some titles for the podcast. Give me five open ended questions and then write the show notes or the event description for it based on this. And so then what it does is it takes all of that. It also takes my booking link. so after we get off the call, I process the transcript two minutes after we're off the call with, with the potential guest. They now have a list of all the notes of what we're going to talk about and the link to book in to schedule the podcast. That's all done. Two minutes. Something that used to take me half an hour to do before no longer takes any time. And after we record a podcast, cause now we've recorded, sometimes it changes. I get it to read the transcript and then pull out the topics, key insights, arguments, conclusions, quotes, all sorts of stuff. People who are not podcasters, you'll find that I talk a lot about the value of recording conversations and transcripts, and I think that it's Culturally, that we will start to accept more that, you know, if you're in a business situation, we're probably going to be recording that conversation because it's too useful not to, frankly. It's so useful to have a recording and then take that recording, transcribe it. And then take that transcription, we can turn it because AI is brilliant at taking unstructured data and turning it into structured data. So what I mean by that is something that's messy and a conversation is very messy, quite often back and forth, a lot of repetition. Oh, I forgot to tell you blah and then AI can turn that into a numbered list with subcategories and categorize the whole thing and make it nice and tidy so you can scan, read it later and make it useful for people. Sometimes I'll use it to process meeting notes like can you analyze the following meeting transcript, give me an overall summary, give me the key insights, the takeaways, the next steps, steps and action items for each person, and then any recommendations, and please ignore the small talk. So again, just same kind of ideas. I have some that take in YouTube transcript of YouTube videos, then suggest titles and video notes. I have ones that work on blog posts, so it helps me to write a blog post. In fact, I've created a lot of little mini chat GPT, custom GPT tools that help me with different stages. So I have one that helps me to start to brainstorm the idea of a blog post. And so I'll give it the basic idea. It'll start to write it out in, in my style. And then I have another one that will take that, and it will turn it into HTML so I can copy and paste it into my blog. It's nicely formatted. And then it will also suggest four or five different titles for me. And it'll also give me some image suggestions, and it'll give me the prompts that I can just copy paste into Midjourney to create the image. So I don't have to go through that step manually as well. So it does all of that automatically. And then what I can do is I can say, now, can you write me an accompanying LinkedIn post to go with that blog post? And it'll do it like a summarized version of the blog post just for LinkedIn. And in that, it'll say, Hey, go check out my blog for the full version. So these are all just little minor things that are really useful, but save 20 minutes here, 30 minutes there, 40 minutes there in all of the kind of the content marketing. And I know not everybody does content marketing like I do. So. I think one of the key things is it really depends on you and your use case. So like only you will truly know what your job is and what, what your tasks are. And so. This is where I think that it's important to understand how you can use ai. What are the capabilities, what are the current capabilities, what are the future capabilities? And then start to think about, okay, in an experiment, how can I use this for me? How can I make this work for the problems that I'm facing, the tasks that I need to achieve? So I, I get it to do all sorts of things. Like I have one which is a prompt, which actually I turned into a custom GPT, which I've made public. Which analyzes a sales call and what it does is analyzes a sales call transcript. It pulls out the client needs and objectives. It pulls out the scope of work based on the conversation, a key deliverables that were mentioned, timeline expectations, budgetary indications, decision makers, identify key decision makers, stakeholders, the competitive landscape, potential objections, and next steps. And it takes all of that and it turns that into a proposal. So you can say, first of all, give me all of that summary of the sales call. And now let's turn that analysis into a proposal. all of this, by the way, stuff that I've created and made available. I've got so many of those kind of prompts that I use on a regular basis. I've got a big spreadsheet that I just chucked them into. And then when I started creating, like one of the things that I've been doing a lot lately is helping businesses to put in place an AI usage policy. So what I did was I created a a little custom GPT. I probably should explain what a custom GPT is.

Louise McDonnell:

I was just dying to ask you.

Alastair McDermott:

so basically chat GPT have made this functionality available where we can upload documents into what we call a custom GPT. It is like a little mini bot that's personalized for you and you can upload your own instructions and your own information. And so I have one, for example, I have one that is the write like me one that I created let's literally call it write like me and it's private to me, nobody else can access it, but I've uploaded, I have seven or eight books, I've uploaded the text of all of those books there. I've also uploaded the entire archive of every blog post I've ever written. So there's probably, you know, 150, 200, 000 words that I've personally written. In the kind of the art documents archival of that and so when I ask it to write like me, it's not just writing the text in my style, it's actually able to go in there and find out, okay, is there something that you've written on this topic before, what's your thought, what's your point of view on that topic, and it's actually able to add that in, and that's really useful because it means that it's much closer to being the real me. Now it still needs me to tweak this. I don't take this and just blindly copy pasted. I still go in and edit these things before I post it, because I think it's essential that you do that. But it means that you're like your first draft, it's, it's no longer kind of a crappy first draft. this is actually a really good first draft. This is like a draft that's 97 percent done. We just need to go in and tweak it a little bit. That's where I think it gets really interesting because now it's much more of, okay, let's just do a sanity check on it and just tweak a few things and then we're good to go. And I think that's the real power of this stuff is it's getting us, it's just making us faster at getting, places. Now, I say all of that, I still think it's super important as humans, as thinkers, that we do some of the work ourselves and particularly that we think about what we're experts in. Because I know a lot of people listening to this are experts. And if you're an expert, one of the things that you need to do, if you're a thought leader is to develop your thinking. And for me developing my thinking, part of that is in writing. Another part of that, by the way, is in conversations. And you know, you might ask a question or say something that triggers something over. I think, Oh, I, you know, I need to, you know, take a note with my pen, which I'm holding up here on camera because people can't see that. But I take a note and think about that. But the process of writing. Is important in developing our thinking because it helps us to formulate our thinking and helps us to develop our thoughts. And I think that we can't lose that. Like we need to, keep that as well. I'm happy to now have more time for that type of deep thinking and that type of writing, because I'm not spending as long creating a LinkedIn post or something else, which is kind of like, yeah, that's basically the same LinkedIn post as I posted, you know, eight months ago, only just rehashed to mention the latest technology or whatever. So like, I don't really think that I need to spend a huge chunk of my brain's processing power On something like that versus actually let's go and finish the book that you're working on. That's far more important for me personally to work on it. and so I think that it's important to distinguish between those two things. And I've created a lot of tools and frameworks around helping people to understand AI and to use AI in business. And so I've created those frameworks. If I want to write a blog post about those frameworks. I am okay with using AI to do 99 percent of the grunt work because I just say, Hey, here's my framework. Can you go write me a blog post about this? And I've already got the framework page to, you know, written on my website. So I'm okay with, AI taking what I've already done. I don't feel like that's cheating or something like that. And so that's a case where I'm quite happy to have AI do the mass amount of work. I think that coming up with those frameworks in the first place, that's where you need a human involved. a lot of our intellectual property, things like that. AI can help support a lot of that with brainstorming and things like that. But ultimately we need a human at the controls, at the at the steering wheel for a lot of that stuff.

Louise McDonnell:

Now, I agree with you. I mean, I think at the end of the day, you said it earlier on, Alastair, you know, you're looking down through LinkedIn, and you're reading the posts, you can spot the ones that have been written by, you know, and they're not, obviously, the voice of the person that has posted it, you know. So like I'm a huge fan of chat GPT, but I actually, I think the prompt I use the most frequently is to check syntax and grammar. So I still write everything myself and I'm gone really super lazy in that when I'm typing it into chat GPT, like I won't even check the spellings if the letters are all mixed up. I'm like, it'll figure it out. I think. Your own turn of phrase, and even if it's not a hundred percent perfect, but it's not grammatically incorrect, it's going to stand out more than something that's, you know, so obviously a chat GPT or Claude or whatever it is. Okay. So where's it going? Like where do you see it going in the short term?

Alastair McDermott:

Well, it's an interesting area where it's going and We we've got these frontier models, which are chat GPT and Claude and Google's Gemini and Microsoft co pilot, I guess. I don't know if they really count as a frontier model anymore. So we've got these guys at the forefront who are doing some really cool stuff. They don't really seem to care all that much about us, the user. Because like the cool stuff that they're giving us day to day. Is actually stuff they like crumbs, they accidentally dropped on the floor as they were off chasing this magical AGI, which is artificial general intelligence, which is basically the idea or concept that AI could be as smart as an actual human being. and so they're, all in a race against each other. And they're trying to be first to create this AGI and they're saying, Oh yeah, here's something cool we made along the way and, giving to us and, even the amounts of money that we're paying them collectively for subscription fees is only a drop in the ocean compared to the, the money that they're actually spending on Google is just bought five nuclear power stations. Like that's crazy. You think about this, you know? The power requirements, the CPU requirements GPU requirements for these data centers. Elon Musk just built a data center in 18 and 19 days which is astonishing And this is just purely for AI. So what we're using is actually just a small drop in the ocean in terms of where the future of this stuff is going. And this is why I think that. If anybody thinks that AI is overhyped, I think you might be wrong because I don't think that all of these people would be spending all of this much money. If there wasn't going to be something in it, we're not talking about huge amounts of money. We're talking about country GDP for 10 years, 20 years, 50 years amounts of money. Like this, it's, it's astonishing. Like we're talking trillions of dollars. And so I think that the smart people at Google and open AI and, Microsoft, like they wouldn't be investing in all this if they didn't think this was going to be big. that's what they're concentrating on. I think that what we and collectively you and me, and the listeners to this, the people who listen to my podcast, I think that we should be doing is trying to get as much value from these tools as we can right now, and also learning how to use them so that when the next iteration comes along, we're going to be best placed to capitalize and use those tools when they come along. Like for example, Anthropx computer use, which I mentioned earlier, that doesn't really work very well right now. But one day we're going to wake up and they're going to say, Oh, by the way, we dropped it, the latest version of computer use Oh yeah, now, now it can like make, you know, 300 people redundant in your organization. It's like, Oh, we didn't see that coming. Well, actually we did, we saw the preview. It just wasn't very good, but that doesn't mean that it's not going to have a significant impact. And like, I think we need to be prepared for the longer term, which is, this is going to make people in jobs redundant. All of us who sit at computers are at risk. And so we just think about that, like the societal implications and all of that kind of thing as well. it's not there yet, man. It's, it's still a long way from that. But there's also an accelerating effect when we get to the point that AI can start to improve itself. That's when we'll start to see this massive acceleration of it getting better at doing things. And that might be a bit scary when that happens. So I think that's why people want to put things like safeguards in place. So yeah, I don't know if there's any of those threads you want to pull on, Louise.

Louise McDonnell:

Well, you know what I want to talk about now is, what's your experience about adoption of AI? Because in my experience as there's a lot of people out there that still look at me and go, Oh, do you use it? And I'm like yeah, like you go down to the schools and, or they're like, Oh, we have to keep the kids away from AI. I'm like, you know, whilst I said I wouldn't employ someone if they didn't know how to use it. So I think it's here. I think you've got to embrace it. But again, what's your experience? What's the rate of adoption in your, estimation?

Alastair McDermott:

Well, there's two different parts to that, I guess. There's the actual rate of adoption that I'm seeing. And then there's what we're seeing in surveys and things like that. I'm seeing it very slow in Ireland, particularly in the West of Ireland, where we live. And I'm a little bit disappointed that it's still so slow. I also think that anybody who tries to get kids not to use it is absolutely nuts. AI is going to be a big part of our education and our workforce. So we need to start learning how to use it properly. not trying to stop the kids using it, you know. And if you were old enough, and I'm not, but if you're old enough, you might remember when calculators are brought in. That teachers protested against kids using calculators and it was all sorts of, you know, I think in the U S they, they went to Congress about it or whatever, you know, they were all good about the idea of calculators and now it's just accepted yet everybody uses calculators in maths class and that's, that's the way it is. There have been a huge number of surveys done on AI usage and I'll give you some quick numbers. One was done in by the federal reserve and Harvard. they surveyed 7, 500 us residents. The university of Chicago worked with statistics, Denmark to survey a hundred thousand workers. And AI surveyed 10, 000 of their workers. And what we're seeing in every single survey is that lots of people are using AI. And so in a kind of an aggregate number, 65 percent of marketers 64 percent of journalists, journalists write a lot, so they're certainly finding it useful. 30 percent of lawyers, which is interesting, are using AI daily. So those numbers are daily. Usually seeing 20 to 30 percent increase in productivity and sometimes more. And in fact, Amazon found that they saved 4, 500 developer years of work and 260 million in yearly savings. Those are insane numbers. 4, 500 developer years. That's not a typo of work. So they were able to, use their AI tools to do a lot of software development and speed up development. So. We're seeing that the numbers are showing lots and lots of people are using AI. And there's this disconnect in business where employees are using AI and employers are not seeing the value. And they don't know that there's a value from AI because they're not being told about it because there's no incentive for employees to tell employers, to tell their management that they're using AI. There's loads of reasons why they shouldn't tell them. You know, there's a fear of punishment. There is kind of being told, Oh, this vague. Oh, you know, you can't put any proprietary confidential information or you get fired and they don't really know what that means. Like they don't tell anybody that they're using their own chat GPT account at work. You know there's the fear of, oh, well, am I going to be given more work if I get more productive or is mine going to lose my job? There's just no benefit. So why would you tell your boss that you use an AI when you don't see any benefit in it whatsoever? So I think that it's important because a lot of businesses are not seeing the value and they're not getting value. Cause if, if you actually got your employees and said, Hey, you know, Start to share the information and, you know, Michael you're doing something with this, go and talk to Mary because she's doing something similar. And if the two of you get together and share your knowledge, maybe we can create a tool or something that will help make everybody more productive and make everybody's job easier. So bring AI usage into the open give people training and give them permission to use it. that's something that, I think is really important I don't know how many people who are listening to this are. Our employers with employees. But if, if you are, then I think you've got to look at normalizing AI usage and reducing the fear of using AI and disclosing that you're using it. Because I think there's too much to gain if you can get people talking about it. There's, there's so much that you can gain. And then if you're a soloist listen to this just start embracing and start figuring out how to use it. Don't use one of the free accounts for 20 minutes, find that it doesn't do something like exactly what you want. You like, you have to spend time. And that's where I think it's important to get a little bit of training and you know, just figure it started experimenting and playing around. Because what you can get from it, if you can get 20 or 30 percent savings on time that's a big deal. And then sometimes you can get more than that. I was working with a consultant who has a training course and she had a gap in the course that she wanted to add. A bonus module to her training course. And we were on a call and we were not intending to do this, but I said, can we just like have a quick go at doing this? And so she wanted to add a module to her training course on behavioral science and cognitive biases. Cause that's what, what wasn't there. And so we outlined the course, we outlined the module. The full one hour bonus module of a course in depth to the point that we exported and we had it ready to put into slides so she could start recording. We did it all in 45 minutes and it was good. it wasn't rubbish. And I think this is one of the things when people start using these tools, they see a lot of rubbish getting output. When you get good at using them and you're able to give it good input, you can start to get really high quality output and that's when things get, interesting. so we created a six module bonus that included exercises for every module. It was going to be 60, 90 minutes long. We had all the slides developed all the talking points, all the titles suggested images for every slide. And we exported it as a CSB. We're then able to import that into Canva to make the slides. And now all she has to do is, And I start the slideshow and start recording her screen and she'll have that recorded total time to have the bonus module done. Probably three hours, two hours, something like that.

Louise McDonnell:

Fantastic. having the content there is sometimes, you know, have it started?

Alastair McDermott:

Now. We were able to take her book and put her book in and we were able to take the outline of the existing course. and then she was able to tell me the names of the experts, we were able to say, okay we want to create a module about behavioral science and cognitive biases, particularly bringing in the work of X, Y, and Z and those were people. and so if she wasn't an expert, we wouldn't have known to put those names in. And this comes back to it, you know, so, and we said, Hey, this is a supporting module. That was going to be a bonus module to go along with these other modules. So it puts it all in context. And I think like, that's a word that I use so much with AI. Like we even call it the context window is where we, we have the conversation. Like what, what can I see? What does it know about it's in the context window? So yeah, if you can put the right information into the context window you can get some really great stuff out of it.

Louise McDonnell:

Yeah, I completely concur. So for anyone that's not using AI, you really are missing it. And it's time to look at it. And, or if you're using it, if you're using a free account, if you're not really thinking about your prompts, if you're not working hard, if you're not getting the output you want, you know, then you just need to change what you're doing because. The capability is there. I suppose that's the key message from this, isn't it? The is there. Okay, well, brilliant. So any last bits of advice then, Alastair, for people going forward?

Alastair McDermott:

I think the most important thing is that you do have to experiment and start to learn. there's all sorts of little foibles and, you know, sometimes you're better off starting a new conversation than continuing an old conversation. So, you know, start a new conversation more frequently than not. There's lots of little things. I've got some resources on my website. I've got like a ton of blog posts about different things about this. Like one of the things I think is important to understand is when you're using what I would call a private AI or a non private AI, and when you're using a non private AI, then you want to be careful about any kind of sensitive information that you put in. Like, for example, you probably don't want to put in the transcript of that sales call. That you recorded cause it probably has proprietary information of the prospective client. You don't want to put that in there, or you don't want to put like somebody's medical records or something like that. So you've got to think about that a little bit and educate yourself a little bit on these things, but the best way to do it is just get started. that's the most important thing. So I would say just, you know, dive in register a chat, GPT account. If you haven't got one already, start playing around with the free version. And then maybe watch some videos or read some blog posts about it. Try some different prompts, try, just make something fun. Like, don't, don't worry about the work side of things for, for a bit, just experiment and play around with it. and then pay the 20 bucks that it costs to sign up for the paid account and just test that as well and you know, see where you might be able to find some value. Think about what are the things that you don't like doing? In your work, are there any of those that you might be able to make faster or better with AI? And if you want help, then you know who to reach out to.

Louise McDonnell:

absolutely. And there you go. Where can people find you Alastair?

Alastair McDermott:

So I'm at AIPowered. biz. So AIPowered. biz. And you can download a free guide to the goal framework, which we spoke about earlier. If you go to aipowered. biz slash Louise.

Louise McDonnell:

Fantastic. And if people wanted to find you out on social, where would they find you?

Alastair McDermott:

I do 99 percent of my socials now on LinkedIn. I don't spend a lot of time anywhere else. I just got tired of it everywhere else. So yeah, I think LinkedIn is the least worst of all the social networks. So that's where I hang out.

Louise McDonnell:

Oh, I have to say it's my favorite place to hang out as well at the moment. So they'll just find you under your name there, Alastair, yeah. That's

Alastair McDermott:

right, yeah, just, just search for me or check out the link in the show notes.

Louise McDonnell:

Exactly. We'll have it in the show notes as well. So look, thank you so much for coming on and being a great guest. I have to say I really love the AI stuff. I find that it has just changed how I work, changed how productive I am. and it has changed. You know, I suppose what's capable what we're capable of. And I think if you're working, even if you're in a team. I have a team of people, but I'm still working by myself because I work remotely. it's company. It's somebody to have a conversation with, which is extraordinary. Like, what do you think of this? Have you thought of this? Da, da, da, da, da. And I, I have those conversations and I'm very happy to have them. So, thank you for having this conversation with me, Alastair.

Alastair McDermott:

no problem. And next time what we'll have to do is, get you your own custom GPT that will have all your instructions and, and all your content in one place. And maybe that'll make you even more productive.

Louise McDonnell:

Exactly, and we've actually started doing that already in our company, so we've actually started creating them for our different clients, so that we don't have to retrain it every time, so we know the voice, we know everything totally love it.

Alastair McDermott:

that's the next level, so, so if somebody, you know, starts to dig into this stuff. That's the next level is if you've got the basic stuff down, go and make yourself a custom GPT and teach it how to be you. And then you don't have to retrain it every time and it already knows you and what you like. And yeah, I think that's kind of level two.

Louise McDonnell:

Yeah, yeah, exactly. Yeah. Yeah. Okay. Thanks very much. Awesome.

Alastair McDermott:

Thanks, Louise.