.png)
The FODcast
In the FODcast (The Future of #DigitalCommerce) we explore the real career stories of the people who have made it to the very top of the sector and those who are working at the cutting edge of innovation and change right now. Listeners to the podcast gain insight into the journeys industry leaders have taken to be where they are today, the challenges they are facing now and their aims for the future.
The FODcast
When AI Takes the Busywork: How Teams May Evolve in the Age of Agents with Rob Goodwin
AI is not just changing tools, it is changing teams, skills and how customers find you online.
In this week’s episode of The FODcast, host James Hodges sits down with Rob Goodwin, Chief Data Officer at MSQ Partners, to talk about what is moving fast, what is hype, and what leaders should fix now.
We cover:
- AI agents for triage, code generation and repetitive analytics, and what that means for junior and mid-level roles
- The rise of AI Ops to monitor, tune and audit models in production
- Why first-party data quality still beats shiny tools, and how to clean and enrich what you already have
- Practical steps to align clickstream, CRM and purchase data so personalisation actually works
- Trust in an AI world, including QA, transparent citations and brand-owned communities
If your 2025 plan touches AI, data or search, this one is absolutely ESSENTIAL listening.
🎧 Tune in now
#AI #DataStrategy #DigitalCommerce #MarTech #Search #Personalisation #FODcast
Simply Commerce is the leading supplier of talent into digital commerce across technology, digital marketing, product, sales, and leadership.
Find our more about our approach and our services within digital commerce recruitment here: https://simply-commerce.co.uk/
Hello and welcome to season seven of the Fodcast. The podcast focused on the future of digital commerce hosted by Simply Commerce. Season seven promises to continue to bring you some of the industry's brightest minds across the globe as we unpick the sector and where it's heading. From war stories to strategy and technology deep dives to future trends, we cover it all as we continue our journey to have one of the most popular podcasts in commerce. Before we start, if you enjoy our content, please do hit the subscribe button on whatever platform you're listening on, like and share on socials. Hello and welcome back to the forcast, The Future of All Things Digital Commerce. Today, I'm very pleased to welcome Robert Goodwin, Chief Data Officer for MSQ Partners to the show. Welcome, Rob. Hey James, good to meet you, pal. Thanks for coming on the show. So MSQ Partners have a global operation with offices in more than 20 locations and sit at just under 2,000 headcount. The impact of AI is a daily conversation for Rob. And today we get to unlock his insights and opinions on AI, from how it's shaping technology teams to the growth of AI-powered influencer marketing and how first-party data is the real gold. Yes, this is another AI-focused conversation, but I have a feeling it's going to be a good one. But before we jump in, Rob, do you want to give a quick overview on your career today and how you've ended up as CTO at MSQ?
SPEAKER_00:Yeah, cheers, James. So when it comes to data technology and insight, I've I've had the brilliant uh background of working client side and agency, consultancy side as well. So grown that career over many years, worked on media advertising in-house, um, and now get to work with God, over 50 to 100 uh global and local businesses. So I get to understand what's going on, what's happening in the real world from B2C, B2B. And the career has gone from hands-on analytics to data strategy to then marketing technology implementation, to then laddering up and then bringing that all together. Where the beauty is is then being able to distill complex areas and uh taking all those terminologies and bringing that to the board as well as then bringing that to C-suites and executives to make better data-driven decisions ultimately and deliver some monetization as well as efficiencies and effectiveness. So uh really interesting, really fun, and it's all powered by awesome teams that sit uh sit underneath me.
SPEAKER_01:You did a far better job at introducing yourself than I ever could have. So thank you for that, Ralph. Uh years and years of experience to uh to pull back on. Um so yeah, I'm obviously looking forward to this conversation. AI is, of course, the hottest topic right now. It is literally pretty much every conversation. Um uh so let's jump straight in. Um, first of all, I think we want to focus around technology teams and the impact that AI is is is having with them. So, how how do you see technology teams now, Rob? And how do you see AI shaping those teams over the next sort of two or three years?
SPEAKER_00:Really interesting because this there's so much evolution that's going to have to happen. We can already see it in the market with God, Facebook spending millions on bringing in AI engineers, for instance, into areas and now actually closing their uh workforce. So they've said, oh, actually, we need to stop hiring some of these areas. So there's a huge explosion of bringing people bringing in at the top levels. Where it's becoming challenging, I think, over the or will become challenging over the next five years or so, is what happens to the lower levels and the mid-levels, because we've got AI coming into play, you've got you know GitHub and co-pilots that control roughly about 40% of all code generation today as well. So all those simple repetitive tasks are being uh absorbed and uh and supported by AI solutions. So this then means do we need fewer people at those kind of um mid and junior level roles? And I think that's where the evolution of the market's going to probably end up going. Um what it doesn't mean is the people that are currently in their roles uh need to worry just yet because it's all about the QA and the usage of those tools. But there's a huge evolution of going, how do teams become smaller, more efficient, but then get superpowered by heroes or unicorns that can do awesome stuff with AI development and software engineering. When it comes to the analytical side, then it goes, but who's QAing this? Who's actually tracking the data or the code books themselves? How is that being developed in the future? So it's an exciting and probably worrying time of what's going to happen over the next five years or so. I think that will stabilize over time, and and there will be some new roles, maybe not as many, but there will be new roles generated from that. Um, but it's the repetitive tasks that are at risks. So that's where you then go. Who is then uh running anything that could be your BAU, simple code task books, or um just any kind of repetitive automation within reporting, within setups, within audits, for instance. You see this also with JP Morgan, they're piloting AI systems that basically triage bug tickets. Uh, so any simple first line or third line um uh requirements that are coming through, they're being pushed via AI. So to either solutionize, if that doesn't work, but then pushes to a human. And this is where people are going to be coming out of the loop, but basically it's removing tasks that we really shouldn't be using human workforces for if we can avoid it anyway. So you then go, um, that's being removed. Challenges though, is how do graduates and junior people come into play and learn the trade and learn the complexity in the future? So that's where HR and training are really gonna have to evolve, going, how do you either fast track people quicker and how do you learn and learn maybe universities or schools in a better way in the future as well?
SPEAKER_01:But that's exactly what I was gonna say, because normally it's these easier, possibly slightly more mundane tasks that juniors cut their teeth into, and that's where they start and then develop. So if we're gonna remove this whole layer of of um of work, then what would a junior developer do? Because you can't just come in and know something. So, where do you pick up that commercial experience from to be able to get to a mid-level to start working on the more complex tasks? Because we what we can't find ourselves in is a situation where in five years we have no junior developers or we have no um third-line supports because we had no first line supports three years before that.
SPEAKER_00:Yeah, and and I, you know, shock, James. I probably don't have all the answers for this one just yet, but what we're trying to look at is firstly get the AI solutions in that work and work not only for us in-house at MSQ and across our agencies, but also for our clients. So it's setting them up for success in the future. So at the moment, we're not seeing it being any reduction in workforce, it's about implementing those solutions, which still takes time, and you hear all these horror stories about POCs failing. POCs are meant to fail in some instances, so that's not such a bad thing, it's about shifting the dial and moving forward, and that's where clients are going. But there are a lot of success stories as well, uh, and implementing AI chatbots or AI customer experience solutions or AI to support HR and finance teams as well in housing, as well as then thinking about marketing and optimization. Um, but where it comes to at the moment is about implementation and setting up, that will then transition from right, if they're in play, where is then the redirection of resource or the cost saving of this? So I think it's about not uh replacing resource that might naturally um natural attrition, so people leaving, um they probably won't be replaced, and going, how can an AI agent, how can the AI code books support that in the future? But then it comes down to training, and this is where we're working with HR internally in other areas of going, right, how do you fast track people and how do you just give them these tasks? And it really changes the dynamic of what you do for shadowing, or how do you then give them specific project bases? More and more development needs to be done here. I don't I don't think anyone's quite got the solution just yet. You can see Klarner went well in front. I don't know if you've seen this with their um obviously the customer service environment, but they've had to bring humans back into the fold because it hasn't quite worked. I think it's the right gamble to take, but you then go, it shows that you have to be more cautious about this. So going, you know, just doubling down on that isn't going to quite turn out uh quite um quite the way you want it. So you need to probably put some more planning into play, have some more uh investment into the structure around it, and what does that and what does also the consumer want as well when you think of the consumer-facing area versus internal like HR finance and the infrastructural setups?
SPEAKER_01:Yeah, the note about climate is quite interesting. Um, I mean, ultimately we need businesses like that to to double down and explore these avenues because ultimately they're the the pioneers for everyone else. Uh, like you said, in this case it didn't work out, but they would have learned a hell of a lot, not just internally, but the industry itself would have learned a lot as well to be able to then use that to then shape decisions moving forward. I think um where I find it interesting is just how those juniors will develop because ultimately there's always gonna be a stream of junior professionals across all industries. And as we start to remove some tasks from them, I can't see a way in which they aren't going to struggle to develop their skill set unless AI becomes part of that development program, but then they're gonna lack the natural problem-solving skills that they might have figured out before we used AI. So I'm not sure I certainly don't have the answers, and a lot of this is well above my head, which is why I love speaking to you, and I'm picking people like yourself's brains because you're far more educated in areas like this than myself.
SPEAKER_00:It's and and I mean, I'm on a few councils like the DMA and uh uh and other areas, and they're discussed, we're all we're discussing this quite heavily, uh especially with graduates and what do you do for graduate training or or non-graduates trying to break into the real world? Do people go to you know the big question is what would you study at university today? Uh what jobs are going to be around uh for future for you know 10, 15, 20 years time place? It's the market will be evolving, um, not as critically, I think, in the next you know, three to five years, but by that five-year mark, I think there's going to be some bigger changes, and that's where yeah, having the good training process, having uh the schooling, setting people up actually at educational level, I think really needs to change. But this is then governmental kind of level we're talking about here of going, how do you teach children and kids when you've got LLMs at their fingertips, and uh and that's going to change the workplace completely. So uh I think that's the way the market's going. Um, at the moment, though, it's about just implementing the solutions and getting people used to using AI agents versus agentic AI. Most people aren't really using agentic AI on scale, agents are falling into play a little bit more. Um, and that is eroding and getting rid of those simple tasks and those autonomous tasks, which in truth, most people shouldn't or want to be doing those. So getting those agents to do that. Uh, but like you say, that that could reduce, say, 20% or 40% of people's jobs day in, day out. So you then go, well, do you need to replace those roles, or does it mean just people need to do more with their time? And I think it's probably going for the latter at the moment, it's doing more with your time using these tools at the moment.
SPEAKER_01:Yeah, it's certainly an interesting landscape. Obviously, we've been here before when everyone feels like their jobs are at risk and turns out actually they've not been at risk, and it's about evolving your role. Yes, there's always going to be some uh um there's always gonna be some jobs that do um unfortunately disappear. It's just natural. Um, but it's a that's why it's so important to evolve your skill set. And I think the one key area that I take, no matter what our role is, is to try and implement the human touch, as we've touched upon here already. You still need that human touch to give yourself a differentiator to an AI, and whether you're in customer service or in engineering or in finance, whatever. The fact that you are a human and you can add that human touch is going to make a big difference to um how you work with AI to enhance your or your offerings either to your own business or to your employer.
SPEAKER_00:One 100%. I think um uh I had I had something, you know it's it's human plus AI co-pilots, right? So, and anyone some some jobs will disappear. There's no two ways about that. That will definitely happen. And then, but it doesn't mean the people in those roles um are are necessarily at risk. It's about how much they employ and engage themselves in new skills, new AI tooling. So, what it does force the hand of is continuous learning right now. If ever it's been more important, everyone should always be doing it anyway, but now more so than ever. My god, everyone needs to be using AI tooling, they need to be exploring it, they need to be researching, this needs to be not only on company time and their own time, and and it's about evolution, it's the future, it's I don't think there's any avoiding it. There's a huge amount of noise though, and there are you know, it's maybe not as good as everyone's uh hope is still, but it's still significantly better. Everyone's had the everyone's on board now to the LLM side, but um the way I think we're trying to describe it is going if you know, think of like Formula One, right? For instance, you go, who's the human driver is say the engineer or the analyst or the setup, and they're in the seat, but the AI is like the pit crew, uh, they're the supporting team around them to optimize that performance, to help with like the BAU setup. So it's going, how can you become better as a whole? And the people that understand not just their individual part of the business but the wider process of it and how AI can then support it, they're the ones that can win. I think this is also where I say roles like architects who don't just um uh design what needs to be built, but um how that in how that integration works, the entire ecosystem, they will become so much more valuable. So it's understanding the wider ecosystem, it's understanding who do you work with and wider than just the tech team, it's working with the marketing team, it's working with the sales team, it's working with HR finance in different areas. So those types of architects and those AI engineers or basically AI infused strategists or AI-infused creatives, they're gonna become the powerhouses, they're the future. Um, so in that human plus AI is that is is the setup. Lots of training and exploration is needed.
SPEAKER_01:Okay. Well, we're already seeing a big increase in uh people looking at upskilling in AI, and there's still a lot of question marks around how, but uh yeah, I think that's it that's inevitable given we're at in the journey. If I asked you if right now you believe AI is going to kill the role of a junior developer in the next five years, what would you say?
SPEAKER_00:Yeah. I I would hate to uh hate to put it on there, but I think it will reduce the numbers you need. Uh it would be what I'd be firstly saying is we will need fewer of them. Uh, I don't think it will kill uh kill the role itself because you still need to bring people through. I think it's just gonna reduce the numbers of Juno developers you will require. What it might do though is change that Gina developer, like a DevOps grew decades ago, I'm sure you would have known, is then it's gonna go into AIOps. So that's where it might evolve into.
SPEAKER_01:Okay, interesting.
SPEAKER_00:People that tune, monitor, and audit the models and the agents or the um uh or or the agentic uh infrastructure for autonomous decisions, that's where that might change. And I think you would have probably seen this with you had analysts that automatically became data scientists, um, not necessarily with the same skill sets, but I think the roles might evolve, there might not be as many, but um I think there'd be a bit of an evolution of what people would be managing and what they'd be doing within them.
SPEAKER_01:AI ops is gonna be the next buzzword then, right? Okay, what I find really interesting here is obviously it's gonna be it's very different, but I have a lot of conversations with leadership professionals, and they have kids that are somewhere between like 12 and 18, and almost all of them are using AI as part of their daily life. And I think something that we're gonna have to get our head around is the fact that there is the generation that's coming through are gonna be far more skilled on a lot of AI-related um products and softwares than people like ourselves, which then ultimately puts people maybe 10 to 15 years younger than us with limited career experience in a more senior role than us when it comes to anything that's AI focused. And I think there could be a real shift in how teams are set up in the future with a lot of their senior AI professionals maybe being between the ages of 22 and 30. Yet there are a lot of other pressure of the business being older than that and their far broader technology experience, but actually in a in a less influenced type of role. Um, I mean, we've seen this with Meta, and as you touched upon, didn't wasn't the guy they offered like a quarter of a billion dollars to like 24 years old, which is crazy, right? Um so lots of uh yeah, I guess we'll see how it unfolds, I guess. Um, but underpinning AI is data. Um, data is absolutely critical to all organizations, but especially first party data. But uh, what it'd be interesting to hear from you, Rob, is why why is it becoming the most valuable area right now? Some say more valuable than gold.
SPEAKER_00:Yeah, and uh or data's the new oil, it's uh it's it's been around for a few years now. It's um and I still uh like wholeheartedly believe it is still the number one powerhouse. If you can own and control good quality data and consumer data, think about the marketing world and customer experience and understanding individuals. If your brand can own that data, it's way more powerful than anything else that you can start to do, especially but only if you maintain it and you have strong brand equity as well, and you've got strong brand trust with those customers. So it doesn't take away from good products, uh good marketing and setup, but that data is a powerhouse for better marketing, better understanding, better optimization. And if you've got this continuous learning environment, data is the powerhouse to do that, and that's where it supports around, I suppose generally it's around contextual personalization. So driving data to deliver really strong, relevant experiences uh for customers, whether that's them coming inbound to call centres and like the Klana experience, that's probably why it didn't quite work out, is because some and maybe consumers aren't quite ready to always be talking about bot. Uh, so there's also that evolution. You had Tesla's coming into play as well with autonomous driving, which obviously whether you can use that term or not, I'm not sure, I'm not 100% sure. But uh people are getting used to they have to get involved and used to that environment, and they're they're uh so there's that market dynamic which takes a bit of time. LLMs happened pretty much overnight, so people's uh onboarding and evolution of using these tools is becoming much more so, and I know there'd be other areas we discuss as well, but being still contextually relevant and creating those great engaging experiences as well as fast experiences as consumers are wanting things quicker and easier, and solutions need to be seamless. Um, the data is the powerhouse for that. So, where are you going? Why are you going to it? Uh, what's your behavioral signals that are powering this? How do you understand uh what do I really want to be served and what am I really interested in? Like South James with fitness and understanding what do we want to go after. You want to go, I want to understand who I am, I want to be served the stuff that's relevant for me and the best content that's that I can engage with, as well as then understanding what's the relevant sub-content, uh, sub-content structures that align to that. So maybe it's about other sports or events or holidays or other areas as well. So um having that first-party data is really powerful. It's not it's on its own, though, you can use contextual signals in the market as well. So think about federated data learning as well. But if you've got um uh inferred data signals in the market, so like what your browsing behavior is, or maybe using these lifestyle or uh locational-based segments like uh that we would have used from uh the likes of uh UGov and Cantor and different areas, but you go if you combine those then with first party data, own personalized data, you can make the best content in the world and you can surface relevant content to your audience. If you're doing that, you're gonna cut through, you're gonna build trust, you're gonna try and create something meaningful for the individual and try and link that back to the brand. And that's how you win in a unbelievably busy and social first world that's hitting us all, and I'm sure we're being drowned with content. I saw a study the other day um uh that was basically saying consumers' attentions haven't changed, we're still consuming content at the same pace. It's not like we're getting uh uh we haven't got the attention spans, but what we've been hit with is at least double the amount of content that we've seen maybe 10 or 20 years ago. So it means that our uh our attention on content isn't changing, but the amount of content being thrown at us is at least doubled, which is crazy. More than that, surely, yeah. And it could be even more, and then you go, Well, okay, but so therefore, what are you looking at? And it just means that you've got to cut through in an unbelievably busy world, and it's why that marketing has become less efficient because there's more at it and more uh algorithmic uh setups is social that's forcing stuff down consumers' throats. So you've got to make sure you're relevant there and using that first-party data and doing data matchbacks and using data clean rooms, super important to get to that one-to-one level, and I think that's changing the media environment and showing brands why powerful it is and how powerful it is to collect that data.
SPEAKER_01:So, how many negotiations? So you've um you know you you speak to brands on a daily basis. I mean, how many brands do you feel as though are where they need to be with their data? Because my experience as a customer, I would suggest probably not many.
SPEAKER_00:Well, yes, and and funny enough, it's I've been repeating this probably for about 15 plus years now, James, and going getting that data quality right. And it's the same with the AI world, getting that data layer right and making sure you understand what you're collecting from your customers or what you're collecting from your business or your HRE finance systems is still number one, and you've got to build those platforms at the beginning. So you've got to get your foundationals right, otherwise, you're just supercharging it with awesome technology and AI, but it's powered on poor quality, and getting that quality right is still the critical thing. I think it's the uh horrendous analogy here is going, if you had an hour to chop down a tree, you'd spend the first uh 50 minutes sharpening your axe kind of thing. Well, that's just data quality. Get your data and collect it in a really possible way first, and it makes the end point so super easy. Problem is the really interesting part is that last 10% the experience at the end point. The problem is you've got to do a lot of grunt work and set up to get yourself there. Um granted AI does make that a little bit easier with getting the data into a better quality in the future, but uh setting yourselves up with then the clients, it's a real mixed bag. Some are much better. Subscription-based businesses, they're doing a really strong, good positioning. So um uh like think of Netflix, they've got this, uh they've got so much data, they're they're collecting the behavioral setup, they've got a really cool 90-second rule. I don't know if you've come across this before, but what you click on in your first 90 seconds helps predict your binge behaviour on what you're doing on that that given time. So they use their data in really clever ways. They're probably at the the the north star kind of level with the Amazons and the other areas for next best action. But more of your mid-sized businesses and your smaller size businesses, they're they're struggling. They they it's where do you focus your efforts, your time, your cost versus sales? And sometimes taking a step back will really supercharge you into the future and dial up that revenue or monetization of your your assets. Interestingly, what I am seeing though, if you're a subscription business, you're doing well, and because you've got better first-party data, you've got something to set up. If you're not, and you're more in like the FMCG space or uh you're a publisher, you're in a struggling area uh because of content availability or direct ownership of clients. And this is where I'm seeing a lot of them transition to more uh customer-first uh subscriptions or trying to find a way to collate free communities or content to absorb that first party data. So I think virtually all sectors and categories now are trying to collect it, some more so than others, and some more successful than others, because it's also not just about collecting the first party data. As a consumer, James, I'm sure you get hundreds of emails in your inbox, you probably don't read half of them. Uh, you've probably bought you know thousands of products in your lifetime, and then you go, well, but why do I want a relationship with a toothpaste, for instance, or why do I want a relationship with my uh cat food for my cat? So you go, it's all about curating that in the best possible way, and it's not going to work for absolutely everyone, but um but it will for some, and there's it's it's working out where's the brand versus the consumer first party data and what do we want to collect and not over overwhelming customers uh by trying to do too much and do it in a privacy set uh sensitive uh setting.
SPEAKER_01:That's the other challenge, and it's like where's the balance between privacy um and performance and getting what you need. Interestingly, we had um a lady on the show at the start of this year, and uh she's from a loyalty vendor, and she was saying some of their best customers uh have questionnaires they send out to their customers to A, get the data and B, obviously make sure they use it appropriately. And like, I mean, obviously, there's a number of challenges that come with that, as in getting the customer to complete the questionnaire in the first place. But actually, as a customer, when I look at the brands that I shop with the most, I would 100% give them two minutes of my time to know me as well as they possibly can do, and then send me everything that's relevant to me, my interests, my goals, whatever it might be. And also knowing that then that would reduce the amount of rubbish that I get come through as well. That just gets hit delete because now it's almost like noise, a bit like what you're saying about marketing on socials. You don't pay attention to it because it's noise. I go in my inbox every day, I've got 50 emails, and I just go delete, delete, delete, delete, delete. I don't even look at what they say. Um, so I think there's a lot of opportunity there that that brands are missing.
SPEAKER_00:It's 100%. Having the brand equity and good quality products is still so important, and and building that brand marketing around it because you've got to cut through, and you'll pick brands that you go, I love this product, or I want to hear from them, or you can see there's relevancy to it, I want it. Uh, they're the and you've just got to double down on that, or you've got to create that ecosystem so that works. We've done something really fun and really cool with the AA, for instance, um uh and in their CRM emails, just putting quizzes in, just like which is all on platform, you can do it within your app, within your mal app, whatever it is, and you can click through on it. And what that did was drove thousands of engagements of just having a bit of fun and just asking some fun questions on what do you think about automotive or um uh mechanics or different areas, and it just gives a bit of insight into the consumer, but it's good fun, and you get to see what you where you write versus other competitors uh or other other uh other customers, sorry, in that instance. So you get to learn something as well. So it's giving back a bit of fun and infers a bit of knowledge as well. So um there's ways to do this in less um direct or awkward settings as well. It doesn't have to take minutes, it could be seconds and And it's just about getting uh good quality, either in third or first party data.
SPEAKER_01:No, I think that's certainly how brands will differentiate over the next couple of years, and I have no disagreements there. Before we move on, if come uh companies have our data, they've been collecting it for years. I mean, look how long we've had nectar cards, right? 20 years or so. Um to those businesses that have a lot of data that is messy, that isn't structured, it can they just buy a shiny new A or our AI tool or tools and make do? Or do they really need to go back into the crux of it, go back through their data, put all of these shiny news or tools on hold for six months, work their way through it similar to sharpening the axe, and then invest? What would your advice be?
SPEAKER_00:Oh, it good, yeah. I mean, really good. And we we deal with this all the time for our clients. So it does depend where they are, and it does depend on the ecosystem and the quality of the data. What I'd always say is uh I think I talked to one brand uh a couple of years ago, and they've gone, they had a few hundred thousand customers, and they went, data's such a mess, we're gonna have to throw it away. And I went, oh my god, please do not do that. That's that's a big decision. Please do not do that. We will clean this, we will get that into a good position. So what I will say is AI has helped us because we use AI agents to look at cleaning, quality setup, re- tagging of data, as as boring as that might sound, getting the good content tags, so creative assets tagged up really well, getting uh understanding of data signals, permissions, how to contact consumers, what have they landed on on their websites, getting that all set up and cleaned between clickstream data from the website, to CRM customer portals and what they purchased. We can use AI agents to help that. Um but if the quality of it was poor in the first place, there's only so much cleaning you can do. You can get it into a format that can be used in these systems, and that would always be the first point. Get it to a good enough state so it's usable. If there's then gaps, then you can either buy, enrich, or start on a data strategy to um uh to plug the holes and go back and ask consumers as long as you've got a relevant marketing reason and a business and customer experience reason to talk to them about that. So that's how we'd normally work in it. Clean it up first, then look at the enrichment gaps. Um and whilst you're parallel tracking that, you can do that on the big bulk data. Moving forward, you can have the processes, the strategies, and the data strategy environments to set yourself up for new customers coming in in the future as well. So those things can be parallel tracked, you don't need to do them all at once, but you can clean what you have and set yourself up for the future so it becomes a lot easier, a lot simpler, and you maximize it. And it's not just about the data though, the data's there, but it's what's the use cases to use it for. So optimizing the website experience, optimising click-through rates, conversions, or looking at better dynamic content for media optimization. It's then going where's the business real focus, where's the monetization of where are we going to use that data for? And are we going to create a new product from it, for instance, like OpenReach, looking at which uh which street level they should be going to to put more fiber optic cables in, for instance. So there's real good uses of this for operational and product development as well.
SPEAKER_01:Okay, thank you. Good good answer. We can almost box up and use that to promote with around what companies should be doing, because I'm sure that's a conversation you have on a daily basis as well, right? So um so look, so we can't go much further than our LinkedIn news feed right now without seeing how Chat GPT in particular, um, but AI is going to drive the Let's State retailers as a prime example of their sales. Obviously, we've seen they've got the the all-in one now, obviously powered by Shopify. Um many would suggest that if you're not AI first in your search results, then you're going to struggle with uh building out a with it with having success ultimately over the next five years. I mean what let's start with this. I mean, what are your first thoughts on AI-powered search?
SPEAKER_00:I I mean, I I know you just said five years there, James. I I think it's oh god, it's evolving so quickly now. Uh five years, you'll be you'll be dead in the water, quite frankly. I think it's in six to twenty four months, you need to be uh you need to be playing it. And the problem and the challenge that everyone will have, and marketers and your digit your heads of digital will be facing right now is it's evolving so quickly, uh, and the algorithms are changing so quickly, uh that it's really hard to stay at pace and stay stay in tune. It's like what do we fix for? So um, however, uh to not scare anyone is that the simple basics of at least web uh crawlable search is still having really good strong SEO, good content, good searchable content, uh and engaging content. So the the algorithms still work on that. However, when you get to LLMs, it's all about uh scrapability uh for AI search, and this is where maybe ghost websites come in or digital twins, as uh as we're calling it, is a digital twin website, but it's an AI searchable website that comes into play, and that's what we've been building for clients at the moment, is going what is what is scrapable, what is easily uh searchable, so that way it can be relevant, and you can get the blue links on the uh uh the AI chatbots uh to fuel it.
SPEAKER_01:But the future really quick, so just really quickly, what what is a digital twin website? I've not heard that before.
SPEAKER_00:No, right. So you bait you basically it's it's the same website, but it's a crawlable website, super, it's a much more easier crawlable website that um is fundamentally designed so that AI chatbots can scrape it. Uh not a it's not a consumer visible uh front-end setup. So it doesn't have the uh we wouldn't be going to it as consumers, so it's not a CX optimized customer experience optimized website, it's a text uh and content optimized website that AI can then scrape, can go through, can refine it, and then they can reference back to it.
SPEAKER_01:Okay, so businesses would have two to they'd have a customer-facing website and a robot-facing website, basically.
SPEAKER_00:That's exactly it. Yeah, that's exactly it, James. And and I you then go, but how do you then optimize that for not only? I know most people probably think ChatGPT, but you've got Claude, you've got Perplexity, you've got Grok, you've got different, you're now not just optimizing for most people used to be just for Google, you're optimizing now for five or six different areas. Generally, they use different um uh scraping capabilities, but they are have got their own nuances. Um and uh but it still comes down to having great content and answering questions. So it's about content choices, answering those cons uh questions, having reviews, having good community setups, um, but also then just making sure it's uh it's engaging uh for audiences if it's still on Google for chatbots, making sure it's readily available and you've still got traffic going to the main site. So that's where having that digital twin makes it easier for them to refine it, search through it, link back to the main site, um, and uh and make sure it's answering questions. I think it's it's that kind of conversational kind of comment that's gonna be really key.
SPEAKER_01:Interesting. I'm gonna uh ask a few more people about that and just get some some more uh more thoughts on the digital twin website. It seems it seems like that's gonna be a pretty standard thing in the next sort of six or twelve months, but uh yeah.
SPEAKER_00:And that's a new new naming for it. There's you've had your your text and your your back-end links that would uh that would have already been scraped in the past anyway. It's just taking it and probably just dialing it up with a uh with a bit more power.
SPEAKER_01:So I spoke to someone in digital marketing not long ago, uh more specifically, very focused around SEO and SEO optimization. And I asked him a similar question. Um your answer was five years is too long, probably in 12 months. I think you said it's gonna be a big, a big increase where now. His response was uh he doesn't see it being an impact for 10 years. The complete opposite end of the spectrum to you. Uh some of the dates, I don't have the specifics, I wish I had them right now, but I think he said chat GPT search is like one billion a day, Google is like 18 billion a day. So there's such a big discrepancy. He also referenced that YouTube has about 4 billion searches a day. So ChatGPT isn't even at the same level that YouTube is, or I think TikTok obviously appreciates ChatGPT has come out of nowhere and got to that point in a very short period of time, but there's still a large customer base that don't use uh Chat GPT, and there's many that probably don't even know what it is other than just an AI tool and that don't believe it's ever going to impact their life. So um this is why I love this show because I love getting people's opinions, and obviously everyone has their own opinions and it's and that's fine, but it I find it really interesting. So I guess just talking to you now about that person's uh opinions on what we've just discussed. What would you say?
SPEAKER_00:It I think where it's probably coming from is in regards to uh if it's around chat, if it's about LLMs driving AI and becoming the the primary setup, yeah, maybe I'd I'd agree with that because uh it might take about 10 years. ChatGPT is the third most uh used search capability at the moment uh after YouTube and obviously Google. Uh and it's probably about four to five percent, as you say, James, it's about four to five percent of total search. That scaling though, remember that's gone from zero to about four or five percent in the space of what a year, two years, and the search capability is increasing increasing. What's probably happened though, and this is where it's scarily moving so fast, is uh depending on when you when you're chatting to him, is that Google AI is now become going to become probably the primary basis for Google search in the future. So if that becomes Google AI, you now need to be at that zero-click setup for Google AI. Uh, and that instantly changes what does your website become? Do you even go to a website in the future? So that's the scary part, isn't it? Yeah. So 10 years for a website build, categorically not. Um, it it will be within the next 24 months that changes need to be made significantly. I I do not question that purely because AI search will be taking zero-click data and your scalability of it will be critical to that. So making changes now is a must. Um becoming uh a not Google being taken over, though. God, yeah, that will take some time and and there will be some needs for uh and they will defend that because they've got Google AI coming through. So whether Chat GPT and the others do take over, they will still become critical parts of it. But I'd anticipate with Google AI and maybe even Apple uh creating their own uh uh AI search capabilities, that will still be important and will take time for um Google to become the leader in the marketplace. So I think there's uh definitely a lot of time there, and whether that even happens, probably not, but I can see ChatGPT and some of these others growing, I can see it probably getting to a 10% level at some point within the next few years. Okay, okay, transfer is not so sure.
SPEAKER_01:So something I think a lot of people are concerned about is the reliability um of AI. Uh is it is it truthful data? Obviously, we know that it can be influenced um by websites, blogs, etc. Um if the LLMs can absorb sort of public web data, then how as a customer are we likely to know what's true versus what's what's not, what's manipulated?
SPEAKER_00:Oh, it's such a challenging one, James, and I face this day in and day out. It's get you know, just buying products off Amazon's you trust the reviews, um, anywhere else. I there is not an easy answer for this one. Um, and what my biggest worry actually is in this area is that consumers don't question what they're seeing, what they're reading and setting up. Unfortunately, that is the mass of consumers globally that they probably don't, they haven't got the time, they uh haven't got um uh don't want to put the effort in to then start going, well, that questioning what they're being seen. So they're they're seeing stuff from either a search bot and they go, well, it's Google. I trust that. Yeah, but Google scraped thousands of articles to get to that. It's not actually Google saying this, or and I think you're right, that is the issue.
SPEAKER_01:You you often go to these places because you're short on time, or you're not you're um uneducated in the area. So you go into an area which you believe is educated, it tells you an answer, so you automatically take that as correct, and it actually we we know that's not always the case, actually, it's really the case.
SPEAKER_00:And and this is where it comes interesting, is because and this is old stats, this is probably um when was this? It's up to about June 25, right? So this this year, but roughly I think it's about 47.9% of citations are from Wikipedia or Chat GPT. Wikipedia by anyone. Uh Reddit across Perplexity, ChatGPT, Google AI overviews, is also another big player. I think Reddit is roughly about 21% uh of AI, uh Google AI overviews at the moment. That will change, I have no doubt. But you think you're being a source from YouTube, Reddit, Qara, um, and Wikipedia. This is where a lot of our content's coming from today.
SPEAKER_01:That's massive, right? That's what top of my head 48 plus 27%. I mean, that's uh 70, 75%, give or take.
SPEAKER_00:It'll it changes for different platforms, but they're big players here. So you then go, but these things can be edited and they can be falsified and they can be fake. You really need to be careful about where the citations come from, and this is where you need people to be uh smart enough and engaged enough to question what's coming through. The problem is it's time, it's effort, and it's education. And then you go, well, where do you go to? What's interesting though is going, I think that that that's unfortunately the world. I think search engines will become better over time, and they you know, Google punishes or punished in the past people that put too many keywords on their websites. There will be AI uh evolution here where they will optimize to maybe non-AI um uh clickbait, for instance, in the future, potentially. I don't know yet. But I I I would have thought so to then optimize consumer output. But where brands can win in this is where it becomes interesting using their customer data, using their technology, is then building that brand equity, building their trust in their customer experience, giving good, I don't know, journalistic, good reviews, building good communities that are trusted, that's what you can rely on. So, like one of our great clients, like with Witch, for instance, that does witch reviews, they're they're in a unique position to be really trusted for their position. Uh so how do other maybe publishers or other brands curate content that you go, yep, if I'm reading this, I trust it. It's been reviewed by an expert, or I trust the brand itself, whether you know it's like Neutrogena for face cream or something else. I trust them because I believe in their science, or I trust them because I believe in what they're gonna deliver for me. That's the way for brands to win in in the content space in the creative world. You then just need to make sure it's findable and searchable, uh, and you're pushing that content out.
SPEAKER_01:I think that's a the um that's a good place to wrap things up, too. It seemed like a nice a nice ending on a few words of wisdom for brands and how they can win. Yeah, love that. Um look, that was uh a really cool conversation. I mean, there's there's loads more questions that I had that I would I'd love to get your uh get your answers on, but uh we're kind of 15 minutes in now and I know time's tight today. So I think let's wrap this up for now. And um, I mean I'd love to revisit this episode in maybe nine to twelve months and just see exactly how the how a lot of this conversation has developed in real life and uh what what is happening uh then versus what we've we've said right now.
SPEAKER_00:Love that, mate. I've got a cheesy uh I've got a cheesy line that we could also leave it on here if uh Okay, gone then. Let's try let's try this as a bit of an analogy of going, look, everything we talked about, but you go, if you can treat AI like an autopilot, great at cruising and supporting that always on uh smooth, smooth riding, but dangerous without a pilot's hand during the turbulence and landing. So if you think about it, human plus machine plus the AI, that's where we need to be getting to, and that's where I want everyone, uh everyone listening to think about training, upskilling, learning, and failing fast. That's that's the way to do it.
SPEAKER_01:Failing fast, one of the key things about all areas in life, right? If you're not if you're not failing, you're not learning. And it's important when you do fail, you adapt, you learn, you grow, and you go again, and then you go through that process again and again and again. That's it. Lovely. Well, a couple of great analogies there. Like that one, I love the Shark and Yaks one early on in the show as well. So uh actually the analogies are like one of my favorite parts of these because everyone has really cool analogies that when you break it down, they're so simple, uh, but they're so accurate. Yeah. Wicked. Well, look, thanks for joining me, and I hope you all uh enjoyed the episode as much as I did. Please do like and share across your networks, and I'll see you all again next time. Goodbye.
