Skip to content
PodSubs
ShowsPeopleGuestsJobsNewsEvents
More
BlogPodCred
Menu
ShowsPeopleGuestsJobsNewsEventsBlogPodCredNetworksSign in
Join nowSign in
  • Shows
  • People
  • Jobs
  • News
  • More
    • Guests
    • Events
    • Blog
    • PodCred
    • Networks
    Join nowSign in

Omnichannel

Omnichannel artwork

AI Expert: How AI agents Work & Used in Outreach, Job Security, GEO, Future of PR- James Shamsi

Aug 7, 20261 hr 9 minAudio
ShareXLinkedInFacebookWhatsAppEmail

Send us Fan Mail Work with James Shamsi: https://www.bigears.ai/business-intelligence Follow James on LinkedIn: https://www.linkedin.com/in/jamessamirshamsi/ In this episode, Dominika Legrand, founder of Human to Human, sits down with James Shamsi, founder of Big Ears, to explore AI agents, personal branding, GEO (Generative Engine Optimization), AI search, PR, and the future of business. Cool, now that's out of the way. Personally, having conversations with geniuses like James is why I love this show so much. We talked about what AI agents are. What we can use them for? How the world has been shifting more towards AI implementation and James's experience around training inside companies to help them adopt AI. How AI talks about you or your company and where it indexes the information from? We also discuss the future of media websites, the decline in traffic, and possible ways for these sites to stay relevant over the next 3-5 years. I had so much fun; hope you guys enjoy it too. Finally, some actually fun and creative ways to use AI to create meaningful campaigns that go viral. Contact info: Bookings & Sponsors: contact@dominikalegrand.com https://www.facebook.com/dominikalegrand https://www.instagram.com/dominikalegrand/ #ai #aiagents #geo #citation #publicrelations #dominikalegrand #jamesshamsi Learn More about the Human to Human programs here: https://www.dominikalegrand.com/programs-1

Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.

Transcript

From the show. Tap a time to play from there.

SPEAKER_00There's there's a lot of like uh media monitoring type tools out there where it's like 24-7 always on where they'll basically have AI agents um that find you know PR opportunities for you. Uh so there'll be like finding thought leadership opportunities, stuff like that. And that's either reactively by looking at you know articles already coming out from similar publishers, and then you you do the calculations of the amount of uh compute time that you're gonna be running these GPUs onto that ask, then the cost of the environmental impact is actually in favor of AI.

SPEAKER_02I'm so happy to see you finally. I I I I really want to like actually poke you, like to make sure you're human. So if I'm coming to London, I will harass you once more to see if we can grab a coffee or something because I would love to have a conversation in person with you. And get to know you, the actual human version of you. I would love that. And um, so I'm glad you're here today. I have you to talk all things AI and not just AI, but AI in PR and AI in terms of how people are implementing AI in companies.

SPEAKER_02I know you have extensive experience around it as well. I will ask you for an introduction, if you don't mind, so that we can kind of put you into context for the listeners, and you can do it better than I.

SPEAKER_00So yeah, all good, all good. Uh thanks for having me on. So yeah, kind of a crazy, weird background in some ways. So I started my career making uh viral content at university with some friends at school. Uh, we did like these light-hearted pranks, nothing like the kind of like douchebaggy pranks that you see these days. Uh we got on Good Morning America, we went kind of viral with it. Um and then I left uni and then moved straight to Los Angeles uh without even visiting, just because my sister was there. Um and then yeah, ended up living there for a few years.

SPEAKER_00Uh built a career in kind of like creating PR stunts and like viral campaigns for brands, stuff like that. Um, and then I ended up launching and starting my first agency, which was SocialIQ. It was a uh PR and growth marketing agency. And then after that, co-founded an influencer marketing agency called Violet, and then dabbled a little bit in crypto, which uh I think most people, especially in the LA scene, have done. And then moved back to London during COVID. And more recently, I just finished uh leading AI innovation at the UK's largest ed tech, which is Twinkle.

SPEAKER_00Um, and I left that a couple months ago to now launch my own startup, which is Big Is, uh, which basically helps you, which yeah, effectively has agents that help you kind of discover different market opportunities and then help you kind of act on them, whether it's for PR or product innovation, that sort of stuff.

SPEAKER_02I love all of that. And in this conversation, the first question that I have for you is around AI agents because you did mention AI agents, and I it let me tell you, like everyone is talking about them, but I think a few people actually know what they are and what they do. So can you explain to the listeners and myself? I I'm a virgin in this. Like my c my client talks about it all the time, like and I'm like nodding and like, okay, I have yet to Google and like look into that. So can you explain to us like a five-year-old?

SPEAKER_02Because we need to know.

SPEAKER_00Yeah, all good. Yeah, it's uh it's definitely the way of the future. So, in simple terms, uh, agents are AI systems that can actually take actions for you and then to do things for you. So normally when we talk about using AI, we talk about things like talking to Chat GPT, and it's a very kind of like one-dimensional, one-off jobs and asking questions. But really, you can have AI systems kind of go and control your computer and actually do whole things for you. So that can be things like controlling your computer to do like full end-to-end research tasks or designing things for you, or even just kind of like your admin admin stuff and like bookkeeping things, stuff like that.

SPEAKER_00Uh so I'll give you an example actually that I um set up for Big Is. So we have a sales agent. Um, and so that agent basically goes out, figures out where your leads are, finds the leads for you, searches the leads, enriches their information, then writes these hyper-personalized emails for them, and then actually sends the emails, tracks the emails output, figures out what did good and what did bad to then kind of like improve itself recursively. And so agents are basically these AI systems that can handle entire workflows for you, not just automating one part that you might deal with, like a Zapia or something like that.

SPEAKER_00It's really about being able to take a completely unstructured task or a task with some structure but that requires like unstructured activity around it to then basically perform the entire task for you. So yeah, it's it's it's it's a pretty wild, um, wild place we're in with the steck right now where you can have that.

SPEAKER_02I love that. And you know, um it it's kind of crazy that you just said so casual, like by by the way, that's what we do. Like basically it would research the leads and then it would reach out with hyper-personalize emails and it would send those emails out and just like basically create performance on what worked and what didn't work. Is it possible that I'm already receiving these emails from people?

SPEAKER_00100%, yeah. And sometimes it's really obvious. Like you you go on LinkedIn, I I don't know if you get this much, but I get loads of um really like uh actually I'll try not swear, BSC uh like messages and stuff like that, where like it's obviously uh a cold message trying to pretend like they actually care about what you're doing in your life and stuff like that. But like all it is is an LLM's picked up some details on like what you're doing and it's trying to force its way to pretend like uh you know it's actually done some research on you. Um so you definitely will be getting a lot of those emails.

SPEAKER_00There's loads of spam going out these days um with people just kind of like using these sorts of out-band agents. So yeah, you'll you'll probably be a victim of few of a few already.

SPEAKER_02I I've already been getting, I think, weekly invites for my show. Like, hey, you want to have so and so on your show, and and just basically almost like skimming the surface of my my show and just like maybe throwing the latest title there, or you know, like whatever, and like no depth in understanding. It's just like because you talked about this title, and then oh, because of that, this person is great for you, and then obviously the follow-up sequence comes, and I'm like blocking the centers, like okay.

SPEAKER_02So I I love all of that, and obviously, we just talked about how to do it really not well. Like, obviously you just said you could tell, and it's like, whoa, it's there's nothing of that depth there that you can tell that they actually give a shit, but they just and we can swear on this podcast, by the way. Don't worry. Um, so and you said like in the PR, the AI PR company that you are now building. What let's go there. Like, first of all, are you currently sending those emails out yourself?

SPEAKER_02And if so, what are what is something that you think is it shows more care, but it doesn't make people look like you're just an item on their list of like how do we find that balance, so to say?

SPEAKER_00Yeah, yeah. So basically what what we do at Big Is is that um as soon as a client works with us, we basically do first full research on them to figure out like what the ICP is, so the ideal customer profile. That's the first part, right? Like figure out the different personas. And often when we work with clients, they actually haven't done that step themselves, and they'll be like, oh, they'll just give a really basic generic answer. So first step is that figure out the personas in your ICP, and then our agents will go and do research on example personas to figure out what are the what are their example pain points, um, and even kind of like within their specific niche in terms of uh their uh professional niche, but also like um within their geography as well, like what's going on in their market, to then uh come up with a um even more personalized outreach strategy.

SPEAKER_00But then when it comes to actually creating the email, the emails are always initially gonna sound a little bit AI written, just because that that's how it is when you first start out. But the part that you actually can kind of stop making it feel like it's uh so obviously AI and get it to be a bit more effective is when you then come in as the human and then you do the edits. And now, probably a lot of people uh listening to this are gonna be like, well, what's the point then of using AI if in the end you're gonna come in and do the edits? Well, the way that you make this work is that AI gives you that skeleton after doing all of the research for you.

SPEAKER_00And then when you come and edit it, a good agentic system like ours at Big Is will basically then be able to learn from all of the edits that you make over time so that the future ones sound more and more how you actually sound. Uh, because right now the problem is that people did the first part, right? They just set it up and then send blasts and then things go. But the way it really works is when you're able to give AI actual examples of how you sound as a human, and you can also do that by giving it like previous historical examples, all that sort of stuff. Um, and so that's where you get the real kind of um uniqueness out of that outreach.

SPEAKER_00Uh, and then when you pair that together with like actual agentic systems doing full research on kind of like whole bodies of um other things related to that uh target. So, for example, as you mentioned uh uh earlier, I think you mentioned that you're getting a lot of the emails where people just kind of like skim like your your last episode, for example, right? A good agency system will do a full deep dive into your career to genuinely understand what's interesting to you, not just kind of like the last part. Um, but yeah, it it you need to basically mix it in with um with that human touch.

SPEAKER_00Also because a uh agency systems are still kind of dangerous as well in a lot of ways. Um they get a lot of silly things wrong. Uh have you heard about the thing that just happened with OpenAI and uh Hugging Face? Oh, it's terrifying. So um OpenAI. First, I'll give some background on Hugging Face. So Hugging Face is a website where basically you can put in like AI tools and you you kind of like will test different AI models on this thing called Hugging Face. So OpenAI, uh the makers of ChatGPT, were testing out one of their new models and uh they put it on Hugging Face.

SPEAKER_00On Hugging Face, there are lots of like tests that you can give to your AI to like have uh be able to compare them on benchmarks. So there's like a humanities last exam is one, and there's lots of other different benchmarks from which you can compare LMs. And so OpenAI used Hugging Face for that. And uh on their newest model, which hasn't been given a name yet, they put the model in this training arena and they told it your only objective is to score the highest possible uh mark on this exam. And so um the AI kind of to an extent misunderstood that.

SPEAKER_00Um, and so what it ended up doing was that it ended up deciding, okay, the best way that I can score a perfect score on this exam is if I hack Hugging Face and figure out what the answers are by hacking it. And so literally the model uh this happened last week, by the way. So it's like right now the big buzz in the in the AI space, where this open AI model uh first somehow gave itself access to the internet. And so uh they it didn't have access to the internet, managed a way to hack itself access, managed then to go on the internet and then steal credential information of like admin accounts for Hugging Face, managed to then get into the back end of Hugging Face to find the answers uh to then give itself the answers.

SPEAKER_00And so that's a really good example of kind of like how agents can like uh misunderstand things. Uh although in this case you could argue that maybe it didn't misunderstand, like its goal was to do that, but obviously the the creator's goal uh or rather the open AI's goal wasn't to do that. Um and so these agents can kind of like yeah, go off the rails quite a bit. Um so that's also worth considering if you're gonna use them.

SPEAKER_02And so when it comes to creating one, like in terms of I know you have your own develop in in within your company, but how does one create an agent?

SPEAKER_00Yeah, you can create really simple agents really easily. Um so for any of this stuff, you don't need to know how to code or anything like that. It will all happen as a result of just prompting. And so there's different levels of agents. Uh, the the the most basic agents are the ones where you kind of put together a task for it in the form of a prompt and then you ask it to do it and you can schedule that task. In terms of how you actually do that, though, um, there's two ways that I'd recommend if you're a beginner just wanting to play around with this stuff. Um, the first is using Codex, which is an app by ChatGPT.

SPEAKER_00Uh, and the second is by using uh Claude Co-Work, uh, which is made by Anthropic, which is uh the maker of Claude. And with both of those, the way that you would make your agent is really simple. It's just you'd have your prompt, uh, you would give your um agent within those either of those two systems access to other like plugins and tools, so like whether using Drive or Gmail or Sheets or Superbase, whatever. And um you're able to then schedule tasks. So you can be like, okay, every day at 9 a.m., I want you to go through my email, uh, highlight XYZ, prep XYZ, add XYZ to a sheet, whatever, and then be able to do all of those things.

SPEAKER_00So yeah, if you wanted to, you could basically have a uh a really basic agent set up within within one hour, honestly. So like basic things like inbox management, your bookkeeping things, uh stuff like that.

SPEAKER_02And what what are the tasks that people use it for, the most most common ones? You mentioned already inbox, um bookkeeping.

SPEAKER_00So we can give Yeah, so the the the people that I know use it a lot for scraping, which which I don't necessarily recommend because you know you could argue the ethics on that. Um but a lot of people use it a lot for scraping, there's a lot of outbound use. Um there's also a lot of um QA use as well. Um so for example, having it go through your website, uh making sure there's no like things breaking in certain areas, uh also even like competitive intelligence as well. There's honestly just so many use cases.

SPEAKER_00Uh the l limitations though that I would say are that you need to consider kind of like token use in all of these, but really anything that you can do digitally and an agent can do. Like I even used an agent to prepare for this podcast uh quite a bit. So yeah, the uh the world's your oyster with agents.

SPEAKER_02Yeah, so you signed me that PDF briefing topic thing.

SPEAKER_00Yeah, yeah, that's actually what I meant. Although I did that as well. Yeah, I'm full of anxiety, so I like to be prepared as much as possible.

SPEAKER_02I'm full of anxiety and I don't prepare so much.

SPEAKER_00Oh, okay. You live dangerously. But um, yeah, I'll give you an example. So, like how how I use that. So um for the listeners, I I sent uh Dominica a um uh a list of things that uh I thought would be like cool topics to talk about, but um I was also like, okay, I haven't done a podcast in a long time. I know that I have a habit of kind of droning on quite a bit. So um I basically took that podcast, uh PDF rather, gave it to Claude in uh what's called a project. So in Claude you can create projects, and projects are like for specific things that you're working on.

SPEAKER_00And uh I gave it a prompt, something along the lines of I want you to role play as Domenica and uh you know, walk me through some of these questions and it gave me like feedback on everything. It was like, James, you're talking way too much, you're going down these rabbit holes. And now saying it, I realize I'm already doing the thing that was it was gonna coach me on, but yeah.

SPEAKER_02Do you okay? But it's you know, it's funny because I'm so unpredictable. I don't know if AI can map that because I literally just like ha, I like that, let's go there. So I'm I'm I'm moved by my own curiosity, and that's so random that I don't think I can even predict that. You cannot?

SPEAKER_00Yeah, no, well, if I really wanted to, what the best way that I could do that would be to download transcripts of every one of your podcasts, stick it in, and then be like based off of that model Dominico and kind of how she handles conversations and flows, which is uh yeah, pretty pretty weird. I won't go fully down there.

SPEAKER_02You can even create that, you know, um AI version of me and Esse if we are having a conversation and you know, model me. Um yeah, but you know what? I I I have a podcast I'm going to be guesting next week, and we did do some sort of a prep call. So, first of all, the whole idea happened because I went over a friend who just started a podcast and we had three hours of conversation, and we're like, Whoa, we have so much to talk about. And he was like, Come to my podcast, like, let's do this in front of people.

SPEAKER_02And then we kind of have like the how the outlines of the topics we're gonna we're gonna cover. But he was like, Let's do a prep call. So he's just like almost as if he's interviewing me, like 30 minutes prep call, and like, okay, you know, I think we can we got this. I'm like, cool. Um, but personally, the way I like to prep, I don't know if that's prepping, but especially like we have the topics, and I have my own, I'm thinking out loud. So I can think out loud and just have a coherent way of thinking about something.

SPEAKER_02That is across time gonna be a coherent way of thinking about something. So long as my thinking is sound and I know why I say what I say, like to me that that just I know like that was a thinking, so I learned that, then I learned this, and so I have coherence. You can throw me into any situation, my answer will be somewhat the same because the core is the same. So that could be adjustments depending on the situation or the person I'm talking to. I can adjust to calibrate towards the person so the examples are more real and like touchable to them, their situation.

SPEAKER_02But it's like so long as my thinking is clear about a topic, there isn't so much you can ask me that I wouldn't be able to answer because I thought about it before. Not so much because I practice. I don't know if that helps, but just thinking out loud and like having quality thinking time is what helps me have clarity.

SPEAKER_00No, yeah, likewise. Yeah. My my prep with Claude was Claude just telling me off a lot, which kind of threw me off. It was basically like effectively just saying talk less, which is ironic because it's for a podcast. So yeah, I don't know how helpful it it was, but I mean it's good with the anxiety and stuff. So, like, you know, if I if I was a kid and I had like extreme social anxiety and stuff like that, I can see things like this being super helpful. I mean, even like, you know, if people are like um yeah, when you're on one side of the spectrum where it's like hard for you to kind of communicate with people and kind of like articulate yourself, I think tools like this can actually be really, really helpful.

SPEAKER_00Uh, but then also you've got kids just making relationships for these AI agents, which is terrifying. So it's a slippery slope for sure.

SPEAKER_02Yeah, for sure. Um, you know, I I want to go back to our topic before before we drift off, but I love that small segue because I also think that we give a realistic application of how we can prep for an interview. Like we are not just giving an actual example of how can that process look like. But you also said something in your awesome prep document, which caught my eyes, which was the fact that you were um training people, like actual in between companies, in terms of how do we how do they adopt AI?

SPEAKER_02And and you have also mentioned that there was some resistance around it. I'm curious to hear more of that and just having you explain to what was the context there, what happened, and in terms of how you overcome that resistance or overcame that resistance.

SPEAKER_00Yeah, for sure. So uh just uh for context, so I've done trainings for like enterprises, nonprofits, all sorts of stuff. And you typically get um quite a big spectrum of like types of resistance to AI, and it's always for different reasons. So some people hate AI just because of like its environmental impact, others don't like it because it's you know still in everyone's creative work, and it's gonna have like economic concerns and it's environmental impact impact meaning. Meaning um, so for example, uh just the amount of electricity that AI uses, water that it uses, and also just its footprint as well.

SPEAKER_00Like I know two years ago, um OpenAI announced that it was creating like data centers the size of Monaco, and that was two years ago, so like now there's way more. So, like the environmental impact alone can can be so much, and then there's you know how it stole every bit of creative work that's out there, and then how it stands to kind of like uh potentially ruin society in terms of economic impact and their own personal job security. So there's always many, many different reasons why you're gonna have resistance in uh implementing AI in your organization.

SPEAKER_00And then you've also got the people that have used AI, maybe when it first came out, and they're just like, what is this nonsense? Like it makes up links, it makes up stuff, it's totally total rubbish, and they haven't really come back to it since. And so they've got this bad taste in their mouth. I'll say most people actually fall into that bucket where they're like, eh, it's not that great, it just ends up making more work for me. And so each one of each, yeah, each of those different types of resistance kind of needs to be met with a with a different approach. So, for example, the first thing that I always recommend is don't approach AI as if just don't try and like uh just sell AI to your team because then it's never gonna go down well.

SPEAKER_00What you want to do is you want to sell the outcome. So, for example, if it's a receptionist that's um uh you know anti AI um because they've got like job insecurity for it, which is a very valid concern. Uh, and we can come into kind of like uh how to mitigate that in a bit, but in terms of how to reduce the anxiety of it and making them more open to it, it's things like you reframe it of like, okay, well, now you can get your evenings back and you don't have to go home and stress about. You know, the email that we suddenly got. Now you can have your evenings back for yourself, especially for like personal assistance and stuff like that, right?

SPEAKER_00And so there's a there's a reframing of like not just, hey, we're implementing AI because senior leadership give us an AI mandate to like, no, we're looking at AI and how it can actually help with our work and saving time and all of that. And then for the environmental concern, a lot of people like to talk about the environmental concern, and it is legitimate, don't get me wrong, but a lot of the time it's unfounded. And so what I mean by that is when you actually run the numbers of uh the cost on your electricity and your carbon footprint of using AI versus doing something traditionally, um, it's really obvious in most cases that using AI is actually more energy efficient.

SPEAKER_00So I'll give you some examples. One example is um um QA, uh Q doing QA of resources, where you have loads and loads of resources. Like uh I work with a client who produces um thousands of educational resources uh quite often. And so with them, you would need uh somebody to firstly come into work with their car, sit at their laptop, have their laptop on all day, and you know, take taking many, many days, if not sometimes weeks, to go through the same stuff, doing car journeys back and forth all the time. So if you actually sit down and you you do the calculations of the amount of uh compute time that you're gonna be running these GPUs onto that task compared to the amount of actual time that you're um uh spending uh on it with AI, then the cost of the environmental impact is actually in favor of AI.

SPEAKER_00Um, similarly, also like you know, a lot of people talk about well, you create an AI image and it uses uh uh half a bottle of water to cool it down, plus electricity. Again, really valid because it is true, but also um you could turn it on its head as well and be like, okay, if I film the same scenes with humans, well, the human needs to drive in, we need to uh have everyone come to the studio, we need AC for everyone, sound editing needs to happen, etc. etc. You add all of these things up, and eventually you see that it not always is gonna be the case where it's more environmentally efficient to not use AI.

SPEAKER_00So that's one of the ways in which you can address it, but you need to actually run the calculations and AI can help you run those calculations. You just gotta hope that it's not super biased. But staying with the environmental concern, sometimes that's not enough for people. They're like, okay, but then the other concern is that um that they just don't want to support these kind of like big data centers, uh, which is again really valid, just to kind of devil's advocate for that point. Um, you know, you've got all these data centers um that literally soak up all the water from the land because they need to use it to cool them down.

SPEAKER_00And the land around that is usually really badly affected. Like if it's farming land, it becomes arid and all of these things. So they are legitimate concerns, and I myself am concerned about it as well. But um, yeah, just on this note of kind of like, is it how to reduce your own environmental impact? The next thing that you could look at is um locally hosting your models, so you're not contributing to the actual data centers, and you can measure the uh the actual amount of impact that you're having on the on the grid yourself. Um, so there's lots of things that you can do there as well.

SPEAKER_00And then there's just smaller basic things like teaching your teams how to work with AI in an efficient way. And so what I mean by that is sometimes you see people go to AI and they might have a simple question, and then the AI will give you like a page worth of response when all you need is like two lines, right? And so then it comes down to giving your team uh training on prompt engineering, basic things like you know, just tell the AI to answer you in a sentence if you don't need a whole page. So there's a lot of things that you can do to like mitigate all these different types of concerns.

SPEAKER_00Um and then, of course, you also have the the people that are actually going to actively sabotage your AI implementation. I've had this happen a lot, to be honest with you, where they actively are trying to make it not work. Um, and that comes out of job insecurity, and you really have to put yourself in the shoes of the other person because most often than not, you might end up actually empathizing with them and agreeing that you would kind of do the same thing if you were protecting your job. And that then brings on the kind of more important questions about okay, how can this actually lead to them still being employed, but you also kind of upskidding them?

SPEAKER_00And so, yeah, it it it depends what type of AI implementation you're doing. But for example, if you're setting up a phone agent, which kind of nullifies a big part of the receptionist job, for example, then it's about upskidding that uh uh receptionist or whoever's taking those like inbound calls to set up appointments and stuff to see how you could, you know, uh upgrade their actual working skills, but also get them to kind of use these tools more efficiently and uh improve their employment further. It's really difficult though. So I I do have to kind of cave you up with that.

SPEAKER_00Uh and then uh just on the last part on this, it's the people that have used AI, right, a couple years ago didn't have that good experience with a lot of them, which is most people. It's about doing uh retraining sessions on kind of like the new stuff that's come out, especially in the like in the last six months. The uh the difference in output quality of these AIs has been wild. So I do a lot of workshops with nonprofits, and uh I had a grant writer in one of my workshops. He was like, okay, I've used AI to help me try and find new grants that we can apply to.

SPEAKER_00But then when I used it, it just made up stuff. And so in that session, what I did was I showed them how to use deep research. Uh for those that don't know, uh on Gemini, Claude, ChatGPT, Copilot, all of them, there's a function called Deep Research where sometimes when you talk with AI, you just want a quick answer, but other times you you want it to really go out and like look at hundreds of sources and then compile this like really detailed evidence-backed report. And so in those cases, you want to use something called deep research, which by the way, you can get on the free versions of all of these as well.

SPEAKER_00And so what I did was I showed the uh the person who was a grant uh grant writer for the nonprofit that hey, you can now use deep research and you'll get all of this evidence-backed stuff, and the hallucination levels are are almost almost zero. And yeah, that really helps. So it's a case of looking at why they're existing, and then you can kind of um specifically approach it in those ways. But yeah, it's a slippery slope and analog.

SPEAKER_02Yeah, I know. And and thanks for sharing all of that. It's hard to talk conceptually about something that is so specific to certain industries and and situations. Like it you're that's what you're trying to do. You're trying to talk in broad terms when when in day-to-day for them it's a different issue. Like, and I'm gonna go and give my own personal example for a client because they have a voice agent AI solution for doctors, and this basically is responding to phone calls and it's trained to respond as the I mean, they know the insurance they accept, like what are the treatments that they offer.

SPEAKER_02So the most basic questions that patients would ask before they book in or the appointment link, they want to have the appointment link. So it would be trained to answer those questions. And obviously, we call it AI receptionist, and the same kind of issue then arises. So their existing receptionist can resist that and say, well, uh now we're gonna deploy this. And what about me? Like, like if if you if that thing is gonna answer the phone calls, like what am I gonna do?

SPEAKER_02And so in that scenario, it has become either um the first line of defense, like so, for example, it could be that you are occupied with a patient, you're missing the calls, or like out of office hours or weekends, like when you're not working. So there's something that's covering your practice 24-7. It could be a use case of that, but also it could have been just hey, like, how about your receptionist is doing insurance refusal works or like things that require almost like more, like it has more added value to what they can do inside of your company or practice rather than occupying with tasks that can be now automated and that frees time to almost like high-value tasks.

SPEAKER_00Yeah, exactly. Um, I've got a good example actually. So there was a uh CXCX team, so customer service, customer experience team, and um they did loads of surveys for their customers, and uh they wanted me to basically see how AI could help them out. And initially, the team wasn't happy with engaging with me because everyone's a little bit scared of the AI person. Uh and so we um yeah, we worked together, realized that they were spending a wild amount of time just doing like translations, stuff like that, and they would get uh all these customer server responses, and um they would never have the time to go through the verbatim, so like the you know, the free text responses that people would leave.

SPEAKER_00And so what we did was we we uh showed them themselves how they can build systems where basically we could have AI kind of go through all the verbatim and then give them analysis and stuff like that. And so their jobs went from being mostly sorry, qu uh surveys and translating them and doing like some basic, like uh you know, one to five like responses analysis, where it's like, oh, you know, 90% of people give a five rating and whatever, you can't really do much with that. You can do a little bit, but not much.

SPEAKER_00So it went from doing that to now um creating reports which actually are able to deep dive into the verbatim and get a lot more actual useful information. So their job went from doing the other stuff to now actually taking that information to specific teams, so like now they can go to the product team and you know, specific other team members. So their job actually changed quite majorly, to be honest with you. Um, and luckily they've enjoyed doing that. There'll be some people that will be like, no, I don't want to do that. My I enjoyed my job of doing the other stuff.

SPEAKER_00Uh, but yeah, the that that example is a really good example because it shows that how AI is actually kind of like allowing you to enter this new higher level of work, which just previously before wasn't possible without AI. So yeah, as you said, it's it's hard to give like these general statements. You have to get like really specific into specific instances.

SPEAKER_02Yeah, and and now I think people can conceptualize, they can understand it better than if we're just like raw terms, like you know, like it can help you. So I think both of those examples help people understand and be like, yeah, you know, actually there are tasks that you're doing that are brain that's repetitive stuff, and and you could use your talents in things that require more of a human touch, you know, versus you just doing repetitive stuff. And I I believe that there are people who actually love it. Like I have a friend who loves numbers, he just loves to look at statistics, and that's like for him, like give me the most boring stuff, and I'm so happy.

SPEAKER_02Like, there are definitely those types of people, but I think I think most people, I mean, in my mind, they they would be appreciated if there is an opportunity for them to grow. And perhaps AI helps them as well to kind of step into uh a different role that's perhaps now is more available for them.

SPEAKER_00Yeah, exactly. It's interesting because I know a lot of people as well, especially on the QA side where they're like they enjoy the for lack of a better word, drone work, like they enjoy it because it's they just get to turn their brain off, get paid to do like some stuff, and they can go go and l live their lives um like outside of work and kind of apply their apply their brain outside of work. So yeah, it's interesting because like a lot of people are like, great, you can take care of my admin work, but it's like no, I like the admin side of my work, so yeah, it can get a bit funky.

SPEAKER_02Um I have a question that is not so related to the the AI itself, but I wonder how do you get into rooms in which you are the AI guy? Like, how does it happen for you?

SPEAKER_00Um yeah, so in terms of uh like how I work with clients.

SPEAKER_02Yes. Like you just they invite you to a company and you're helping teams and you know how do you get into that to so that they trust you with that?

SPEAKER_00Yeah, so um when I first started uh my role at Twinkle, it was my first like foray into AI, and then I ended up leading the AI innovation team. And then since I left after that, um I was like, okay, a fun way to kind of like uh get some clients for this is just gonna be by hitting up other people who are in a similar role to the role that I was in, because I know exactly the pains that they're struggling with. And so that was like that was like my in. I was like, yeah, I did the same sort of problem that you were doing, and so yeah, managed to get a few clients like that. And then uh a lot of them also with the nonprofit work that I do.

SPEAKER_00So uh I do a lot of workshops, stuff like that. Uh, and a lot of people uh will just hit me up being like, hey, we seeing that you're doing these trip types of trainings, can you come in and kind of like help our organization with this and that specifically? Yeah, it's been mostly through that stuff. Uh I don't make too much of an effort to kind of like put myself out there for that stuff right now because of focusing on big ears. Yeah, in the beginning days it was just hey, I did the same sort of thing as you did, want to share some insights and then quietly uh subtly sell myself to them.

SPEAKER_02Oh no, I I think that's very organic as well then. Okay. I've just I was just curious, like, how did you come too close to the circles to okay? So we just proactively did that, and then the workshops made it very easy for people to be like, oh, I actually am curious about the same thing. Okay. Were these workshops online or in person or both?

SPEAKER_00Uh all online. Um yeah, I don't really do too much in person. Uh I I like my house. Like staying indoors.

SPEAKER_02I feel you my introverted friend. Perhaps that's why we we couldn't couldn't meet in person because that would require you to leave your house.

SPEAKER_00Yeah, just keep my digital twin everywhere.

SPEAKER_02Maybe there is a time when you're gonna send your hologram to a coffee with me. And you're like, hi.

SPEAKER_00Yeah, that will happen in the future.

SPEAKER_02For sure. Okay. So when it comes to resistance, what you were kind of essentially saying is that there's just ways for you to help reframe what is that for them, empathize with them, but also focus on more of the outcomes, like how things are easier or better or more efficient so that they have that helps them come around of that resistance. Is that what you were saying?

SPEAKER_00Yeah, exactly. And you can, you know, um get get them excited for like the stuff that can come out of it. Like the the the survey example, right, where where they can get excited. But it really depends on the person that you're not gonna be able to motivate everyone to use it. There would always be just some people that are like, I don't want change, I don't like change, go away, leave me alone. Um and yeah, with those ones, you can't really do too much. Um, but you there are a lot of options out there still for most people.

SPEAKER_02I I want to venture a little bit around PR because we did kind of touched on that at the beginning of the episode and just you talking about the personalized outreach um emails and kind of talked about that. But I I also know that with the company that you are building now, there's much more than just outreach and just wanting to understand what are the tools that AI is now providing, especially for PR professionals.

SPEAKER_02And if you could talk a little bit about that, that would be lovely.

SPEAKER_00Yeah, no, 100%. So there's there's a lot of like uh media monitoring type tools out there where it's like 24-7, always on, where they'll basically have AI agents um that find you know PR opportunities for you. Uh so there'll be like finding thought leadership opportunities, stuff like that. And that's either reactively by looking at you know articles already coming out from similar publishers and then being like, oh hey, here's like this insight that I can add to this article or a contrarian viewpoint, stuff like that. But then you've also got other types of uh AI agents that are more proactive, where they'll be like, okay, we're not gonna look at what's already happening in the news.

SPEAKER_00Instead, we're gonna look at what the organization has in terms of assets. So, like, what unique insights and data could the organization have? Um, and so our agents, we give them the data that the organization has, but we also go a step further where we get the agents to proactively think what insights could the organization have that they don't yet have, because most most of the time companies don't have that stuff ready to go, like they don't have the the uh the surveys and the case studies and stuff like that. So the other side of the agents are the ones that are like, hey, you should have these insights.

SPEAKER_00Can we talk to your product managers or whatever to get this stuff to then create these opportunities for you? So it's a mixture of like these reactive uh PR agents and then these proactive PR agents. Um and then yeah, it can go a lot deeper as well, like into how you use this always on market intelligence to then feed your product innovation as well, and like all of these other things. But um, yeah.

SPEAKER_02When it comes to insights, because I want to kind of go back to where is the value for from a perspective of a publication? Let's let's assume Forbes. I'm gonna give you that example because that's everyone kind of knows it. So when it comes to Forbes, can you walk me through a process of what those insights mean? Like like do I like is there value for the value is in something that they don't possess, right? So if you can give them an insight that they don't have, then you're like, oh, actually that could be a valuable article for us to publish.

SPEAKER_02Or how does it look like just for the beginners who don't understand?

SPEAKER_00Yeah, so for for any publisher, right, it's it's about traffic to their site ultimately. And so the way that they'll get traffic to their site is by getting interesting article pieces that will drive that traffic either um organically, in the sense of somebody reads the article, shares with someone else, or you know, it's written by someone, and maybe that person interviews an influencer, and then we were getting that influencer to share the article, it's driving some traffic, or the more modern version, which is um you're writing content that gets cited by uh LLMs that can then drive traffic to you.

SPEAKER_00So for the publishers, uh the main thing is um either it's interesting articles that can get that organic distribution, or articles that will get the citation, and the articles that will get the citations are the ones that will hold usually very niche-specific um insights. Um right now, um a lot of the citations you you'll you'll see a lot of people talking about uh um GEO, which is generative engine optimization, which is kind of like the art of getting AIs to kind of like well, yeah, cite you in the answers so that you get traffic.

SPEAKER_00Yeah. So right now it happens to be the case that um most citations uh are being taken from like listical sites, sorry, listical articles and comparison articles, like buyer guides, stuff like that, uh, versus like six months ago, maybe a little bit more, it was the case that if you asked AI a question, it would like look for a press release and then it would uh quote that, or it was a really heavy reliance on Reddit, which there kind of still is a little bit. And so for publishers right now, um, yeah, it's the holistical stuff, but it's changing a lot.

SPEAKER_00Um, what was your question today answer, actually?

SPEAKER_02No, um, so I was asking um like the process of like how the insights become valuable for a publication, but I think you kind of did that you said either it's niche specific, like there there is some sort of value in that, or it gets like citation through through AI. And I'm curious about it as well.

SPEAKER_00Yeah, the the the citation stuff is is pretty dangerous and scary as well because it's it's so easy to manipulate. Um and so you'll see actually there's already quite a lot of uh fake publications being created, there's fake uh journalists, there's uh fake studies, uh fake influences, there's everything fake now, to be honest with you. Uh it's already happened and they already are successfully gaming the system where you can create even like a um a substack post as like a comparison guide, and um you just basically include yourself within the mysticals and then you start ranking on these LLMs.

SPEAKER_00And so it is kind of a a scary space now where like nothing online is uh it's hard to know what's what's really truly written by an actual uh legitimate publication and what's just written to like fool an LLM. Um and then you've also got like the other side of it, which is how a publications using AI, um, which is really interesting because a lot of journalists, most journalists hate AI, rightly so. The ones that do use it um are using it for like kind of helping them with their research.

SPEAKER_00But the interesting use cases I've seen are journalists that are using AI to identify what conversations aren't being had. So they'll use like always on monitoring services to see okay, everyone's always talking about this, and then the AI. Helps to spot gaps that they could talk about. Um but yeah, it's it is the Wild West right now in terms of kind of like GEO and publishers and all of that.

SPEAKER_02Um you know what came to mind, and I'm curious about this, but you kind of you kind of answered it in the the most negative way possible. Um but I so what about and I think this is more of an intentional approach rather than the how can I manipulate my perception and like all of that. But you said that knowing that AI is like and people use it to AI, sometimes they look people up. Like you can ask, hey, tell me about James, like and just you know, the big ears, or you know, so people look up you look you up on AI, like that's a thing now, especially if you are a public persona or you have a business, like that's a thing now.

SPEAKER_02And uh even though you just said that there are ways you can manipulate that research and like you can plant articles to be like you're amazing and whatever. But what are some of the more ethical ways that you can do that? And I'm curious to see like what are the platforms that that ChatGPT or other LLM is using to kind of source their information so that we can be more intentional around publishing there more of our stuff, not to scam anyone, but to make sure that we are present in the AI searches as well.

SPEAKER_00Yeah. The the ethical way to do it is um with earned media, which is basically when you um, you know, get like a publication to uh the the them get a publication to basically talk about you, whether that's by like you pitching yourself to them and stuff like that, where it's getting a real journalist to talk about you uh or a real influencer or or you know uh real entity to talk about you in terms of what the AI is gonna pick up on. It's gonna pick up on the earned media from the publications, uh so like, for example, if a business insider interviews me or something like that, but it will also use a lot of Reddit, um a significant amount of Reddit, even Reddit posts that are like many years old, it will pick up on.

SPEAKER_00So one of the services that we run on Big Is is GEO services. And so we've run at this point hundreds and hundreds of tests to like look at what's actually being cited, and it will be genuinely like years old Reddit posts. And so if you are, for example, doing like brand reputation or you want to make just make sure you're coming up in results. What I always recommend is um think about what are the prompts that your ICP is gonna ask, run those prompts in a temporary chat. It's really important they use a temporary chat because otherwise the LLM will cater the answer to what it knows about you.

SPEAKER_00And if it knows that you work at that company, it's gonna recommend your company. So you want to create in a temporary chat, put in prompts that your ICP will use, and then see what's coming up organically and then pitch those things, whether it's a publication or a subreddit, go on that subreddit, participate in that community, leave kind of like your thoughts and stuff like that. Um, if you try and game the system too much, especially with Reddit, like you can get caught very easily. So you know, I wouldn't recommend uh recommend that. The right ethical way to do it would be to kind of like look at what's currently being cited and sourced and look at how you can organically insert yourself into those conversations.

SPEAKER_00Even LinkedIn, by the way, is being picked up more and more, we've noticed. So LinkedIn posts, even if no one's liking it, like it's still adding brand equity in a way that it's hard to measure, but um, you know, it will have some have some effect as well.

SPEAKER_02You know, I uh I also noticed that LinkedIn is getting quoted as well many times, even if I have no likes whatsoever, like that's actually pulling information from LinkedIn. Interestingly, Instagram it did not pull my posts and stuff, um, and Facebook either. So somehow LinkedIn was the one that I found personally that when I was posting there, if I asked about myself, like that's where it got the references and my podcast as well. So all the podcast episodes, um, the transcripts and the topics, because it kind of give a general image of what are you talking about, like what are even my website.

SPEAKER_02I think my website, my podcast, and LinkedIn, like these are the uh personally that I've seen um LLM's citing and to have a coherent picture and trying to explain like who you are.

SPEAKER_00So that's yeah, it will change for LLM as well. Like uh Gemini Googles one, um, more readily picks up on YouTube because it's part of the Google ecosystem. So it also varies a lot based on that, like perplexities, um generally the most diverse in terms of sources. Um, and then each we can get really into it. Like each one has different ways of working, like um chat GPTs, how it works is that when you put in a search term, uh or rather, when you put in a query or a prompt, it will then create a search term from it.

SPEAKER_00And you can actually see it. You can, you know, if you if you want to know a doubt about it, you can right-click, click inspect, and then go into the network data and you can see the prompt that it turns it into. Um sorry, the search query it turns your prompt into, and then it basically then takes that search query and then searches that through different providers. So, like most typically Chat GPT is using Bing search, um, and so it will just put that search into Bing and then return what comes back. So yeah, it's a very, very like um, what's the right word here?

SPEAKER_00When something's uh nascent stages, I think is the right word. It's a yeah, very much in its early stages of of the industry. And um, in the future, the way I see this kind of like improving is that uh right now, obviously, these are gameable with listicles and all of this stuff, but in the future, the the the right way of doing it would be an AGI that can kind of pick up on like all of your review data, all of your follower, just basically anything that feeds into your brand equity to then kind of like uh come up with the recommendation. But um, yeah, we're probably not too far out for that, maybe like two or three years or something.

SPEAKER_02Awesome. You know, there's something I wanted to just touch on and not go super deep into, but I know that this kind of perked my interest in terms of you know how you can now get information just using AI, and even now like Google is giving you that AI resume and you don't even have to go to the website. Obviously, that's affecting traffic to the website. So help me understand this like even when it there's a citation happening, if you now know the information, why would you go with the whole thing?

SPEAKER_02Right? Wouldn't that be a user experience that's kind of be like influencing how we are cost to be content moving forward?

SPEAKER_00Yeah, so uh most publishers have found, yeah, like everything's dipped quite a lot. Um in terms of brands, uh, brand traffic, at least for the brands that I've been talking with and working with, it has dipped. Everyone's traffic has dipped. But the publishers have got it all worse. The brands have found that while it's dipped, they're still getting quite a bit of site traffic because typically when you're getting cited uh in an LLM as a brand, you're getting cited because you're either at the exploration or the consideration stage of the funnel.

SPEAKER_00And so, you know, if a brand if an AI tells me it recommends your brand, I'm still gonna click on your company website to go see it myself and see what the product's like. So it's a little bit different for brands, but yeah, for publishers, I don't really see how they're gonna come out on top of this. I mean, OpenAI's made some deals, you know, with like publishers and stuff like that, but they're never gonna be able to recoup the amount of kind of like ad revenue that they got before. Um, so they need to change their business model quite a bit, and we might end up seeing that the entire industry kind of slowly, slowly, or maybe not so slowly, moves towards like kind of like a substat kind of a uh an approach where journalists have like their own followings and their own kind of like uh distribution channels.

SPEAKER_00Because the other problem that you have now as well is like um, how do you as a publisher ensure that the person writing for you is the real person, especially if you're a Forbes and you're just you know selling out Forbes Council things because for those that don't know, you can become a Forbes writer yourself or you just pay them I think$1,500 a month or something and you're on the Forbes Council. And so a lot of these publishers would argue don't really care, like they're they're just trying to make money, uh and so yeah, it's they're gonna degrade in like um trustworthiness as well, and the LLMs will recognize that, and then you'll have this kind of like showdown of who's the most reputable author.

SPEAKER_00I know you say you don't want to go too deep into it, but I'm just gonna add two more minutes of stuff on this, which is the the way that I see this going is that um let's say five, ten years even down the line, right? When all the publishers have pimped themselves out and you can pay to play, you can create these artificial accounts and make yourself a writer and stuff like that. The only way that I see this working is you know how with LinkedIn you can verify that you work with a company if you've got a domain with them. The way I see it happening is there'll be something like that where you can have like this centralized profile for yourself and you can kind of prove affiliation to like the university that you went to, the company that you work with.

SPEAKER_00And then in effect, it kind of like how on the blockchain you can see where everything relates to like one body, like one wallet. It effectively kind of becomes like you can see all of the content that you've created and how that's been spread across the internet so that the AGI, the Advanced General Intelligence, can then see okay, James has posted 10 articles this month, uh, and he's officially affiliated with XYZ Company versus Corey has posted 5,000 articles a month and has not validated their emails.

SPEAKER_00Um and so once we're able to do that, then the LLMs can more predictably kind of like have more trust in what's being said about one thing versus the other. So I think we'll move away from kind of like publishers having the uh the weight and the and the trustworthiness signals to like individual journalists. Um even that's not fully clean because I could lend out my profile to you, but even if I lend it out, you could be like, hold on, James works in petroleum engineering. Why is he writing an article about cryptocurrency or something?

SPEAKER_00So I I do see some hope, but the hope comes out of like the industry needing quite a big substantial change.

SPEAKER_02Yeah, for sure. No, I love that. Um so do you see how like personal branding will be bigger than? Because in the scenario, that the faith and trust will be in the brands, the personal brands themselves, rather than the publications.

SPEAKER_00Yeah, no, a hundred percent, hundred percent. Um, yeah, and even within that, right, you can imagine like as as imagine ultra smart intelligence, right? It'll be able to audit your social media following, make sure you haven't just kind of bought a bunch of stuff and all of this. So yeah, I think I haven't actually it didn't click to me until you just said it, but yeah, personal branding um is likely to be kind of like one of the biggest things that it looks at for determining if something's truly credible or not.

SPEAKER_02Yeah. I I think it's already happening in terms of auditing your social media. That's why the tools now can tell you fake follower accounts, like that's that now there are actual tools for that. So that's why it's so baffling to me when people are buying followers, because to me, like, what's the point? Like if there is a serious brand that wants to work with you, that the first thing they're gonna do is run your profiles through some sort of platform like modash.io. I I use that to audit profiles, but I'm always shocked to see how many people are still doing these things and not and not getting away with it.

SPEAKER_02And like, dude, don't you think people are smart? Like anyway.

SPEAKER_00No, it's wild, yeah. And then they they they set up full fake engagement groups and stuff, like it's a sleeper slow. Like it's it's it's it's wild out there.

SPEAKER_02Yeah, I think we we are trying to you and I also like obviously work smart and build smart. Like, I think that's definitely something to to take away from the episode, but also be mindful of the things that require some actual work to you know, some actual effort to build you know trust with people. And I don't think that you can necessarily skip that. And that's sometimes organic as well.

SPEAKER_00Yeah, exactly. I mean, especially if you're a brand using AI like publicly as well, like with your content and stuff like that. Like you see a lot of a lot of brands just kind of like outsourcing creativity to the AI tools now. Um you see it, like I don't know if you're seeing it a lot, but I see loads of infographics now on my timelines everywhere since ChatGPT came out with the image generation stuff. And it's like when it first came out, I think people thought it was cool and like were paying attention. But now, like as soon as I see an infographic, I'm like, you didn't put in the effort to make this, I'm not gonna put in the effort to read it. And you just kind of like skim by, and that's like slowly eroding trust as well.

SPEAKER_00But to make it a little bit positive as well, I've got a good example, which is um the World Wildlife Foundation, so the WWF, they um they hired a designer that uh used AI to make a campaign. And usually, whenever nonprofits use AI publicly, everyone hates it. Um it never goes down well. But um, this was an example of like how it was used really well. So the designer basically wanted to show how specific types of farming were destroying specific types of habitats, and so what they did was that they um I'll send this to you as well after so you can see it visually.

SPEAKER_00But what they did was that they um got um there's tuna fish farming, and tuna fish farming was really kidding a lot of turtles, uh, and there was also like something to do with um uh some sort of noodles, but palm oil, I think it was palm oil farming that was killing a lot of lions. Um, and so what this WWF designer did was that he used AI to basically create an image where there was a tuna can that was opened, and inside you could see the tuna, but the tuna was made to look like a like a turtle, and so it was like a really good example of like, okay, you didn't just outsource your creativity to AI, you had an idea, you communicated that original idea to the AI, and then it created the content.

SPEAKER_00And so uh that and so yeah, they they made a turtle with a tuna, and then he also made um actually I don't know if it was a he, it might have been a lady, um, but they also um made a bowl of um noodles in the shape of a lion, uh, and then above it was some text about you know how that type of farming was hurting that type of animal, and that went really viral and did really well. So it's a good example of like how if you are a brand and you do want to use AI to create content, uh you know, you can do it in these sorts of ways, which doesn't just erode trust in your content, but it actually can be like genuinely engaging and interesting.

SPEAKER_00And that went viral for like good reasons.

SPEAKER_02I love those. I have to see them myself, so I can just actually but I will put that into the episode as well. We can edit the that post as well so listeners watching on YouTube can see. Um, you know, I think I am I'm I'm 100% with you. I think what endlessly entertains me is Ryanair, their Facebook page. I think they do like acts and Twitter and like Instagram, they're everywhere. But they're also using AI to kind of like their whole marketing is they're leaning into like everything, the fact that you have no window when you're buying your seat, or the fact that you are, you know, there you should be charged if you are wearing jeans, or we're gonna charge you if you're clapping for the pilot after landing.

SPEAKER_02And like, you know, now with the whole door incident that happened, like I'm pretty sure they're gonna come up with something, but they also use their graphics with AI to make it more funny. So it becomes this whole thing where there's someone creative using AI to just make the whole brand just funny and engaging and fun. So I think those are applications are almost like there is creativity in creating with AI versus you're outsourcing your thinking and you want it to do something uh without you actually prompting it properly and giving it actually your consideration.

SPEAKER_02The same goes to content as well. There's a difference between you writing with AI and you don't just just dumping like, hey, write me a post and then whatever comes to you just copy-paste it, versus you having your idea or even edit like initial script or something you wrote, and you give it to AI to improve it to make it more punchy, or to compress it in a way that's to the point, or sometimes to find words that you could verbalize yourself and like, oh, actually, that's a really good word.

SPEAKER_02I could use that. So that again, there's a difference between how you are using it. You can use it to improve your current content and and creativity, or just like outsource it. I think that's where the difference lies. Now, the last thing I want to ask you before I let you go, and hopefully my camera won't die, we're gonna pray for it. Battery life is around one hour and 30 minutes, so that's our max. So we were still good. Um but I think in terms of when I'm drifting towards uh humans, and uh you're not aware of this, I'm building uh human-to-human as a body of work, and many of the things that I teach around humanity and how we show up and relational intelligence, and a lot of that, even in business, how do we relate to people, a lot of that I talk about.

SPEAKER_02And so I think one of the things that you mentioned in your awesome prep document is that there are things that you notice that uh AI cannot replace. And I would like to invite you to help me understand what you meant by them, just based on your experience, especially for someone who created viral videos back in the days.

SPEAKER_00Yeah, for sure. So my stance has actually changed on this in the last week. Um and so I'll show what. Yeah, so you know, still is the case that true human creativity is still winning. Um, and what I mean by that is like right now, at least on my feed, I do see a lot of like viral AI slop content. Um, and it's usually like gimmicky stuff. So, like, for example, um uh I don't know if if this is still going around, but it was like um these like kind of like horny fruits talking to each other and stuff like that.

SPEAKER_00It was a while. Hopefully you've seen that. I just don't sound crazy with that. But there's there's always like these gimmicky like AI slop stuff that goes viral, but there hasn't been many cases at all, actually, that I've seen where it's like truly viral content that AI's made um originally, and it's not just going viral for the sake of it being gimmicky and slop. And so that that's something I still haven't seen. However, what has happened in the last week is that um I've been working really closely with my agents. So um every day, it sounds wild saying it, but every day I have meetings with my agents now.

SPEAKER_00Um so I've got my chief growth officer, and I've got like uh I've got many different agents that I meet with, and uh I've been coaching them more and more. So uh we've been creating like different landing pages to test for our ads. And uh every time I talk with it, I get it to log like our meeting notes and I get it to go to its system prompt and then save learnings about uh what we've talked about and kind of like help it to like recursively improve as a result. And what's happened now uh is that it's actually been able to help me come up with landing page ideas that I wouldn't have come up with.

SPEAKER_00So one example and and that are actually doing well. So like one example is that um uh and I can send a link to you as well so you can check it out, is there's a landing page that we're testing right now where um we basically have uh like message pop-ups come up on the page um when you first go on it to illustrate the pain point. So like one example is when you go on it, it will come up with like a message from the CEO saying, why aren't we on Forbes, question mark, question mark, question mark, angry face emoji, uh and like stuff like that to like kind of in a fun way call out the pain point without having to like say it in all these different ways.

SPEAKER_00That's an example of a landing page idea that I wouldn't have come up with myself, but so far is actually performing pretty well in the test that we're running. Um, and so you know, typically I would have said that okay, AI is not really being able to compete with true human creativity, but in this case it did better than me in being able to come up with this actual landing page idea. Uh, but it only did that after this kind of like continual feedback that I've given it. But not once did I say anything like this idea to create, it came out with that itself.

SPEAKER_00What I did with it was um I have a I have a learning and development agent who's responsible for making sure all my other agents are continually self-improving and learning. And so one of the things that it it it it does is look at like other landing pages that are doing well and stuff like that, and then seeing how we can contextualize and repurpose it for ourselves. So, you know, even now you could still make the argument that, well, they only did it because it maybe you you you copied it off someone else you don't really know. That could be true. 100% that could be true. I just don't know. But it's one of those cases where it was kind of like an aha moment where I was like, genuinely now my chief growth officer is good enough to like be like level to level with me in marketing.

SPEAKER_00So yeah, my stance has changed on that since I shared that doc with you.

SPEAKER_02You know, the the question was what is something that human can still do and what should and now you said, Well, never mind. I thought I was creative and then I got something better.

SPEAKER_00Okay, yeah, it's it's it's a weird world.

SPEAKER_02But can I just say, like, to me, this is fucking hilarious that I can just imagine your vibe. Like, okay, so darling, I'm gonna have a meeting with my agents, you know, and uh we're gonna have a discussion and just it's just you. And yeah, yeah.

SPEAKER_00No, I do it a lot. Like I when when I'm walking my dog, uh, I literally would just have meetings with my agents and they'll walk me through like what I need to pay attention to, and they'll be like, right, change this, change that, and then like they just do it. It is weird, but it's like it's pretty great.

SPEAKER_02And you know what? Someone like he was like, We're having a blast. Like, you know, they have sense of humor, they're fun. I don't know if they do, but or they're just great.

SPEAKER_00I haven't given them personalities yet, but it's it's been on my mind that I want to. Like, for one of them, I told them, like, whenever you give me the uh the daily updates, give me a gif. Um, the the only gif that it ever sends is the one of uh there's a dog and a like a cartoon dog sitting down on a on a on like a chair being like this is fine and everything's burning around it. That's the only gif it ever sends me. So either it thinks I'm doing a really bad job or or I just need to work on it a little bit. But yeah, that was my attempt at giving it personality, and it just apparently thinks everything's falling apart.

SPEAKER_01Oh I love that for you.

SPEAKER_00It's like you created your own little world and it's yeah, it's tragic, but it's it's good and helpful at the same time.

SPEAKER_02But you know what? I think it's very authentic as well because you are creating a company, it's called Big Ears, and and you're deep into the AI stuff, so you you better be embodying it to the core, even if it's like gets getting weird and whatever work.

SPEAKER_00Exactly.

SPEAKER_02So in in terms of big ears, and in terms of what are the things that um people can work with you and what calibers and can you help the listeners uh orient themselves a little bit before we wrap up?

SPEAKER_00Yeah, yeah. So um if you want to use AI to kind of like help you find this PR opportunities, improve your GEO, find more leads, activate them, um, also just even you know get a better beat on what your competitors are doing, uh, so that you can, you know, be far be faster than in innovation, stuff like that. All of that stuff, big is can help you. You can go on big ears.ai and check it out. Um and also if you want to talk about AI training for your team and things like that, uh, or you just want to say hi, hit me up on LinkedIn or or Instagram. Um LinkedIn is is typically better.

SPEAKER_00Um and yeah, you can also hit me up on email anytime, just jameshamsy at gmail.com. Um and yeah, we'd love to also hear what you guys are working on as well. It's always interesting to hear what others are working on and kind of like sharing thoughts as well.

SPEAKER_02Yeah, and maybe someone will hit you up and be like, I need an AI agent team so we can have those daily meetings like James have. Maybe that's the that's what they are desiring. Just say, you know, entrepreneurship is a lonely journey, and you're like, no, I'm not anymore.

SPEAKER_00Exactly, exactly. Yeah, still sounds weird saying it, but yeah.

SPEAKER_02No, it's fine. I love it. Thank you so much for coming.

Open the transcript file

Hosts and Guests

Nobody is credited on this episode yet.

Credit the people on this episode

Be the first to rate this episode

More episodes

All episodes
  1. The Psychology of Relationships: Why the Same Problems Show Up in Business and Love| Christa Coetzee

    Nov 13, 20261 hr 5 minAudio

    Send us Fan Mail What if the problems showing up in your business relationships aren’t actually that different from the ones showing up at home? In this episode, I sit down with Christa Coetzee, Human Connection...

    Credit the people on this episode

  2. The Psychology of Creating Desire for What You Offer

    Nov 7, 202617 minAudio

    Send us Fan Mail If you want access to the training mentioned, go here: https://www.dominikalegrand.com/balance Whenever we have something we think is cool, it's easy to think other people also want it too. Especially...

    Credit the people on this episode

  3. The Psychology of Outsourcing Difficult Conversations to AI

    Sep 8, 202617 minAudio

    Send us Fan Mail In this wonderfully improvised couch recorded episode, you will learn about the newly present phenomenon most of us are facing at every single level. AI text messages breaking up with you. AI firing you...

    Credit the people on this episode

  4. Why Coherent Thinking Naturally Creates More Influence?

    Jul 24, 202634 minAudio

    Send us Fan Mail Have you ever been presented with an idea where you went: "Mmh, I don't quite see the throughline of how this person came to their conclusion." What about a conversation in which someone presented you...

    Credit the people on this episode

  5. The Hidden Power of Expressing From The Right Energy

    Jul 8, 202617 minAudio

    Send us Fan Mail How many times have you said something, written something, published something, only to reflect on it after thinking, " Eh, this wasn't my wisest move. Instead of getting into a shame spiral, there...

    Credit the people on this episode