Skip to main content
Podcast

What agencies need to know about generative AI, with Roy Murphy

By April 4, 2023April 25th, 2023No Comments

Welcome to Episode 84. If you’re interested in the fast-changing world of generative AI, and how clients and agencies are using it, then you will love this conversation. Roy Murphy is the founder of Synthetic, an agency specialising in emerging technologies.

Roy shares with me:
– the main pain points that generative AI is solving for clients, and how he’s currently advising his clients.
– what you need to know right now as an agency account manager about the AI industry
– how he thinks the AI landscape will evolve in the future
– and why agencies should be looking to partner with AI experts to help position themselves one step ahead of their clients.

There are not many people in the agency world who have such a deep understanding of emerging tech, so I was very lucky to grab his time. Please do get in touch with him if you’re keen to discuss any aspect of what we covered.

If you’d like to talk to me about agency account management training, I’ve just updated the website to include three of the main courses that I run to help you retain and grow your existing client relationships. (Account Kickstarter, Account Booster, Account Accelerator)

I’m also launching shortly a ‘managing difficult client conversations’ training. If you’re an account manager and you’d like to be notified when this goes live, please sign up for my weekly newsletter here.

 

Transcript:

Jenny Plant  00:02

So today I’m delighted to be talking to Roy Murphy. Roy is the founder of Synthetic,  a Digital Innovation Agency who helps clients with emerging technologies, such as Web3, AI, apps, voice, chat, AR, VR, etc. He also happens to be the host of the tech and business podcast, Conversations Worth Millions, where he interviews experts in marketing, business and tech, and shares insights and practical ways to win with emerging technologies. Roy, a very warm welcome.

 

Roy Murphy  00:37

Thank you, Jenny, very pleased to be here. Big fan of the podcast. Looking forward to getting into it.

 

Jenny Plant  00:41

Thank you so much, that means a lot to me. Would you mind, you have such an interesting background Roy, We have obviously had a pre chat and you explained to me that you are a creative director and you were a digital strategist. You have this really eclectic mix of experience and skills. So would you mind starting off by just sharing a couple of minutes about what you did prior to Synthetic?

 

Roy Murphy  01:05

Sure. and on top of that, we’ve got to add amateur boxer and failed boy band member. Yep, we’ll get into that later, I’m sure. So I guess I’m a classic T shaped person, I’m just interested in lots of different things. In the digital sphere, because I have got a bit of a background in development, and as well as design and business. I wanted to try and work along all those sectors, and then delve deeper into each one. So that’s what I have done over roughly 20 years. Both on client side and agency side, is to have those many hats. I have enjoyed all of them, I can’t deny I have been sacked out of a few because I tend to not master things, but get into things quickly and then get a bit bored and move on to something else and want to try something else. But I think that’s a good thing. I think that wide interest, as long as you can hone it and have an interest and delve deeper into the correct areas, I think is really interesting.

 

Jenny Plant  01:58

Very interesting. So tell me with Synthetic, I think you’ve been running that now for I think over nine years now. What problems are clients coming to you with? And how do you help solve them at Synthetic?

 

Roy Murphy  02:14

Sure. I think it’s changed over time. Interestingly, because we are really always at the forefront of technology and technology changes, right? We’ve seen in the last 10 years, a couple of waves. Obviously, we have had mobile. So when the clients were coming to us for mobile, and “can our site be responsive”? and “what can we do”? And “why isn’t it working on X, Y, Z browser?” and so on and mobile phones and etc. And we did that for quite a while and that was quite successful. Then we will move on to IOT, and chat and voice and do the same thing. So the question has changed over time, but ultimately, the output was always the same – how do we help our customers? Can we reduce costs? Can we generate revenue?, and that hasn’t really changed no matter what the technology is , because ultimately, it’s a catalyst. It is not the tech itself, although it is very shiny and new and quite exciting. But ultimately, what does it do for the customer, and for the client  and their business? That is the key thing.

 

Jenny Plant  03:10

And my understanding is Roy, that you tend to work upstream with clients, i.e. with C suite level decision makers, who ultimately, as you say, are looking to solve business challenges and business outcomes. So I suppose some clients want the shiny Penny, or they’re being distracted by the shiny Penny, they just want to talk to you about the stuff. But fundamentally, can you explain how you work with your clients? I think you talked about your innovative workshop process?

 

Roy Murphy  03:42

Well, I think it always starts with education, particularly i the newer parts of technology we work with. Often it is “what’s the key pain point that they’re working with?” and that is usually along the lines of either automation and workflow inside, or talking to customers outside. And really, it is those two things always. So a typical engagement would often work within a workshop framework. Okay, there is  a technology that has come out of Generative AI, for instance, in this case, but previously, voice or chat, and IOT. And that would be okay, what do we need to know? What don’t we know? And make it simple, because obviously, you have got C suite, and they are not in there to listen to a whole bunch of information around very deep and technical language models or GANS, or whatever it might be. But what are the outputs of those things, I think it is just keeping it simple. And that is it really, I think, as agencies, our key job is to understand the technology and the consumer user behaviour, and then to translate that for our clients in a simple and effective manner. So that really tends to be just talking and building that trust and understanding what is going on in the market and having a good broad view of the trends and the analysis and what they might affect in terms of their customers or clients and being interested to be honest, when half of it is in new engagements it is just conversation. So these might be In their networking events, or it could be at lunches or it could be slightly more – more kind of random. But the conversation is always the same thing. It’s not selling. It is just-  what are the pain points? Here’s what we’ve seen. This is where the direction of travel is, and here’s some thoughts perhaps on what might be the first steps to take and another conversation builds from there.  So I think it is always really a trust based scenario. You are not selling, you are just discussing and talking about where are we are now? and  what is happening? What are the opportunities? Where the interest is, what the challenges are? What are the pitfalls? That really goes across the board. Of course, when you are in C Suite, you have to make your colour punchy, and you have to make it snappy, no one’s got time, the have got a million and one things to do. So you have got to, you have got to make it quick. So I think that simplification comes from knowledge and expertise and trust in your own understanding of where it is going, and what is actually happening in the market and translating that for our clients. I think that is the key thing and that has never changed whether it’s C Suite or further down the chain.

 

Jenny Plant  06:00

There are not many people, given the fact that generative AI, particularly, has become pretty mainstream so quickly since the end of last year, there are not that many experts, I would say, that would be able to have the breadth of experience and expertise that you have to have to be able to navigate them through what is happening right now. So could you start by sharing your thoughts and perspective on how generative AI has exploded in the last few months? And what, principally, for marketing clients? What do you think are the key things they need to be aware of?

 

Roy Murphy  06:43

Well, it’s interesting, isn’t it because “generative” as a phrase probably was not in the vernacular at all, or in classrooms three, or four months ago just wasn’t being used. Now? – of course, it hasn’t come from nowhere, AI itself is  50 or 60 years old back in Turing times. But it’s interesting, from about 2014 onwards, if you have been in that space, there have been leaps and bounds, not commercially, but under the hood, with all the different kinds of language models in Ganz, and adversarial networks, all the kind of work going on underneath it. And often what will come out in the last few years would be a better quality image generation. And you’d see these little kind of,  ticks up over time, and so on. And that’s exploded over the last year or so with various different technologies. So, over the last few months, really, people have been starting to ask what is generative? And how does it work? And what do we do with it? So from a business perspective, and from a cultural perspective, I guess, the Open AI,  where they launched in December, last year, with Chat GPT, what they have really done as we said, is, they have basically launched the killer app for AI. So we’ve been talking about this for seven or eight or nine years in our business. And that’s been through various iterations of, you know, chatbots, and voice. And that’s using some conversational IPs  as intensive classifications, but those things have now increased exponentially with the amount of big data going into these large language models. And that’s made the whole thing useful. And the reason it’s useful is because, obviously, the data going into is quite high and quite a lot of data. But it’s now just an easily commercially understandable proposition. It’s Chat, it is something easy we can actually use and work with. And that has been a big difference. So really, what’s happened is and if you look at Open AI in particular. There are other models, by the way, it’s not just Open AI and GPT. But as we talked about it is the huge consumer pushing, this is really, I think, right to say, it is probably the biggest consumer application launch in the last 10 or 15 years, it really is and I think the last data we heard, it was 100 million users and that was back in January. Now, interestingly, since then, they have, well would not say closed ranks, but you can see what’s happened, the whole thing’s exploded, everyone’s realising it’s an interesting space to be in. So the data coming out of companies like Open AI, which is understandable, you know, they are a business, it is a little bit a bit less, a bit more closed. So we don’t know what the actual numbers are, but they’re pretty high. So basically, since that point, what’s really happening is we have got the, I guess, the combination of consumer behaviour, technological change, societal change post COVID, all coming together in a very simple interface. So all that technology in the background that has happened in the last 10 years has now exploded. There is a way to get involved in it. And now you have got two challenges as far as we see it, you have – what do I do with it? On the other side of the coin you have-  what we call infinite canvas, which is  “well, you can do anything”. So there is a balance really of understanding how technology works, but also what should you do with it. When you have a massive what’s the opposite of whitespace problem? What do you do with it? well we didn’t do anything. Actually,

a lot of the questions we get from our clients are – well, that’s great. But what do I do? What is it? Do I start with social? And shall I be doing more? So I am using it to transform my social and improve that automation process. Also do I use it for ideation and new product development?, Should I be using it for SEO?, then helping my keyword search and social and so on? So many different ways you can take it. And that’s actually part of the challenge of, of what we’ve been doing over the last few years by re-visiting the last six months of actually helping our clients to understand which of those are the most appropriate? How we will help them best? What they should do first?

I guess that’s how we’re kind of approaching it.

 

Jenny Plant  10:48

Lovely explanation. Thank you. Can you give us an example of, let’s say, for example, a client has come to you with the SEO question. I suppose, when we are having this conversation we are wearing two hats.  Roy, you are talking to your clients all the time, about how they can use it from the marketing client perspective. But also, at the same time, agencies, traditionally, have been leading their clients into the future. So there will be agencies listening, possibly feeling a bit threatened by what’s happening – if I’m an account manager working in performance marketing agency, and one of our services is SEO, I need to be on top of things to understand how I can guide my client. Because we are all learning about this at exactly the same time. So perhaps,  you could give an example, or a use case of how you are helping to guide clients around one particular area? Any area that comes to mind?

 

Roy Murphy  11:54

There’s quite a few, and actually interestingly,  there really is a lot of use cases here. And we have been quite busy. So, a lot of conversations coming back and forth. A couple of them might come to mind. Definitely on the social side and the content side.

There is a quite obvious use case across lots of elements, which is, if you have got content, can you repurpose it? Can you sweat the assets as it were? What can we do with it? Can we use things like GPT, to say, Okay, well, we’ve classified this data and this information, and we’ve got it coming, run it through a couple of automations. Try and maybe something a little bit different, and then repurpose that content, through social. So that’s one thing that we’ve been talking about quite a lot with our clients.

And that is quite powerful. Because in terms of again, this is not to jump in, SEO agencies or content agencies are constantly doing great work. You always need a human, by the way, human in the loop is still very important. You can’t do everything AI –  human and machine equals winners, we like to say, but we have had lots of conversation, that’s probably the number one use case also, I have got lots of content, can you help me to create more content from what I’ve got? Because I’m a bit concerned about generating fresh content with automation? Is it going to be any good? Is it going to hallucinate, as they say, is going to be correct, factual, etc. So there’s definitely still some human interaction there. But that’s quite a big part. And then what you do with that. There are several automations that we are running within integrations, that are helping to use that content, and to repurpose that across different social channels. So that really, is I would say, one of the top three use cases is content generation, by content generation from repurposing perspective, not necessarily from a starting from scratch perspective, although I would add to that, but in terms of research and insight, tools, like GPT, very important, and almost become- and I am loath to say –  they are invaluable,  in our view and you can work without them, but I don’t see why you would. Because, from the creator and the hacker and maker mentality, which is very, very popular and kind of overground these days, trade to  economy. That’s really come through into this consumer application view of AI, with Open AI, it is actually anyone can do it. You do need to know what you’re doing with it. But you can automate it quite quickly. And you can test it quite quickly. And you can use it for research and insight. So it’s not just banging out some quick prompts in Chat GPT. It’s what data have you got? What are you trying to do? What personas should each one imagine? And using it in the right way I guess. That’s definitely one use case. There are definitely a couple more, within automation flows. So much more of what are your internal processes?. What is your CRM doing? And how can we potentially automate that with generative? And again, not just chat GPT but other elements too. How can we automate that. So it might be that your workflow, and your funnel may be- email to customer, email sequence, into database, pull up a database with sentiment, review sentiment, contact customer in certain way, for instance, like nice and simple kind of serial flow. That’s also achieved, but with generative. So these workflows are, so you’ve got the social and content piece on one side, but on the other side, these workflows are also becoming, I think, table stakes with businesses and businesses that are not using them, are wasting time and effort from their staff, doing manual tasks that don’t need to be done by a person, they need to be overseen by a person for sure, absolutely right, I totally buy that. But that’s something we’re seeing more and more and we are working with two, one in particular, but two large enterprises doing exactly that, around their internal communication to start with and again, you can imagine, and rightly so, it’s a new box of tricks, call it that, for want of a better word, it just is. That is the reality of it. So there has to be some checks and some balances within this. So a good way of doing that is to keep it, if it’s business crucial data, you might want to keep that in house. And we will be putting everything through a publicly available, or potentially publicly available, or a publicly trained model if you like. So that is one thing to be aware of. But that is changing because other API’s and at enterprise level, things you can do GPT are actually solving that quite quickly. So that’s another use case is workflow and automation. And there is a lot of ideation as well. So social, obviously, content, workforce motion ideation, in terms of product generation, is interesting. We have had several conversations about that, and how that actually connects with other technologies around emerging tech, such as Web3 but that’s another conversation. These things are all coming together, by the way, this convergence. But ideation is really important. There is a lot of work that is often the same way. Fashion, there is fashion and another industry I won’t mention, which are all often at the very forefront of technology, you can imagine the other one. Fashion is , always, the same with Web3 actually. But the same with generative. We have seen that. In Fashion Week, a couple weeks ago, lots of brands started to do okay, this is generative, how we’re going to do this? And big brands, like Nike and Coca Cola recently, actually, are starting to do what I would call real UGC, which has been the Nirvana for many years, right? Very popular 10 years ago, it didn’t really work, co-creation is fine to a point. But you need skills, you need to be able to have professional skills to do it properly. And it tapered out a little bit, a lot of bang into co-creation. But with new tools? So Coke are doing I think, quite recently a toe in the water. So they are sandboxing it and keeping it very much on brand. But using some of the newer tools like Midjourney and Stably fusing the image tools to get creators to create and then use that as part of their campaigns and be showcased in places like Piccadilly Circus, and Time Square on the digital billboards. Amazing. So my way of thinking would be, and these are big brands fashion, of course, you kind of understand that. But these big brands like Coke and Nike as well and various other ones are understanding that this technology is quite game changing, and they are willing to put, well even if it’s called PR dollars for one of a better word, terrible phrase, but you know, even if is that, it is still happening. If they are willing to back it, you should be too. And that is what we’re saying to a lot of our clients, they do not have to go all in for it. They do not have to say right, we’re doing nothing but AI and generative and GPT, this is great. Nothing else, we are going to sack all our staff and we are just going to use Chat GPT. Absolutely incorrect. What you should be doing is experimenting, and learning and being involved and talking to the right people and so on. And getting those resources together. There is plenty out there and you would be bonkers not to, frankly, because  Schrodinger’s cat as it were is not going back in the box, right, it is out now. It exists. It isn’t going back in the box. There are checks and balances. It is interesting to note that the UK Government, I think it was yesterday , their AI Regulatory piece, has gone very light touch. Some other countries around the world a bit heavier on the AI, and actually, interestingly, not to be political, but I think it is as interesting as part of yesterday’s and the days before Zoom conversations. AI was at the centrepiece of a lot of what was being said, which I thought was fascinating, probably the first that I have seen that because UK of course is, I think, after The US is, I think is second or third in terms of EC funding for AI. Of course, we have got, Silicon Fen and London, etc. and other places too. So I thought that was interesting. That’s actually happening, as well as the light touch regulation, it is actually a bit of a boon for, and a direction of travel for where it’s going from a political perspective. And that’s kind of what I said throughout –  just to wrap up that bit, sort of the long answer. But where I’m going with it is, is that we believe this is not just technological consumer behaviour and societal change, it is political change as well. So you have got the whole four and that’s why we think it’s a massive opportunity.

 

Jenny Plant  20:42

I think that was so well explained. Thank you so much for making it sound semi- simple to follow. Interestingly, I have been following what’s happening with Coca Cola. My understanding, as you said, is that they’ve signed an agreement with Bain and Open AI, and again, your thoughts and advice around, keep experimenting, keep learning, I think is very well placed, because you just need to keep an eye on what’s changing, because Coke is usually very innovative, isn’t it? With the NFT space, I think they worked on Decentraland, they launched a range of NFT. So they are well known for being at the forefront of technology and what’s changing? I have got a couple of questions to ask, just for people listening, that perhaps are not as up to date with what’s changing. Just a couple of quick questions. Has Google said that they will penalise any content? Do you know? Content that has been generated by AI? And the second question is, How reliable do you think is the data that’s being produced? Or the answers that are being produced currently, with chat GPT, which is now four, I’ve signed up for four. Just for people that perhaps aren’t as up to speed.

 

Roy Murphy  22:14

So Google, interesting, and I saw something interesting this morning about Google and my understanding is around their adversarial AI teams actually burying the hatchet to work together. Because there are deep mines and I think you are right in saying there are two separate teams. GPT is a massive threat for Google. It is just, this is such a  bonkers thing to say, for a company that has, publicly launched their latest version, I would say in December of last year, to say that Google are running scared. A little bit is actually, I think, factually correct. So I guess you could look at it two ways. Now, we have spent quite a bit of time in the content space as well. So I know quite a lot. I have done quite a bit in, and around the content and blog space and the AI stuff and what Google actually looks for and doesn’t. And the general consensus is they are going to penalise because from a business perspective, they are a massive competitor. And if you look at, and of course, if listener’s don’t know, OpenAI is I think, nearly 50% owned by Microsoft. So big tech has not gone away. This isn’t a scrappy startup that’s jumped in and said, look at this amazing thing we’ve got. It is Microsoft who invested I think 50 billion or something, something large into OpenAI. So they come in after Microsoft, and Microsoft, of course, to close that loop have been using GPT4.  And they will be doing deals with Apple. They are back in the game, as it were, you know, Microsoft, brilliant company, and very, very influential and forward thinking, but they have not been on top of that big tech tree for several years as we know, the big social platforms and big tech as we know it. That’s changing. I think. So if you asked me the question, I think we have not seen any penalising today, but that’s going to change. Because of course, you have the high level, massive big tech with a 50 year, 100 year plan, where is this going? We can not let this these guys take our lunch, we can’t, we are just not going to. So that’s one thing. That is a really important point to note. So I think it will, because their algorithms will definitely favour Google , always does always has done that is their business model, that’s fine. So that will change it. Maybe it hasn’t today. Things are getting clever though, as well. It’s very interesting, because, of course, if you’re generating content from GPT there are plenty of ways of using tools to recreate that and it gives GPT itself to personify talk, like I’m a real person, dumb it down a little bit, make it less verbose. There’s lots and of lots of conversations I have had about prompt engineering, and so on. So there are ways around it. But then Google have got massive teams that are going to change. So I think I would not be putting out a complete, site or whatever it might be of just generated content, I just wouldn’t. And that’s probably not what it’s for, doesn’t mean to say people aren’t doing it and I know a lot of schools kids certainly are. Yeah, you know, and who could blame them? Oh, can someone write this Shakespeare? This is going to take a while, okay, I get that. So, yes, it is coming. So as I say with a human in the loop never going to change, in my view, for certain things, absolutely automation, but that creativity is going to have to be passed at some point through editing at some point and that might be automated, it might be that, Copier in itself might change and how people interact with it. But it gets back to that co-piloting, that this isn’t taking your, well, it’s probably going to take some jobs in fairness. But it’s not taking everything. And if you’re embracing it, and you leaning into it, and you’re utilising it to make your job easier, I don’t personally see what the problem is, that’s from my side. So I think that is quite an important point to make, actually, that AI by the way, isn’t evil or good. It’s an inanimate object, it doesn’t exist. It isn’t sentient yet. Another point to make and sorry to go back is when we talk about artificial intelligence, and again, I’m not a data scientist, but we work with plenty of good ones. This is basically about huge amounts of data, not about artificial sentience. This is just lots and lots of money, to put behind data, to give it a prompt to what to do next, therefore, to sound more like a human. That’s basically it. There isn’t a box of tricks, it isn’t magic. It’s just lots of data, and transformed and pre transformed. And so that’s first the first, what was your second question?

 

Jenny Plant  27:05

Before I go back to the second question, and I am just bursting with questions, just for people listening, again, Roy, who perhaps aren’t as kind of up to speed or they are catching up with what it is. I think we didn’t start with the definition of AI. But you’ve just said it’s essentially a huge amount of data, who actually inputs that data? Where are they getting that data from, in order to then generate the responses.

 

Roy Murphy  27:35

Just, in essence and again, I don’t work for OpenAI, but, you are basically looking at the Internet, in essence, that is what you’re doing hence, the issue you have got is what is the information ? Are those sources verified? We all know, and you know yourself and the listeners too, that there will be sources, every day, that you will go to, you will go to your trusted sources, whether that be X or Y, whatever your favourite is, and they are probably where you get most of information from. So if you are leaning one way or the other, you are probably going to get the data. That’s, you know, we all know, what and where has this come from with, you know, various social platforms, you’re being served up more of the same thing, and that’s advertising anyway, you know, based on your profile and your usage. And so there is a question of bias, to be aware of, as well, of course, and it is a big challenge, because the world’s data is being put in, or what the online data available, has been put into these large language models. But if that data is flawed, well, then the answer is going to be flawed. We have seen that in many cases, particularly some of the biggest social platforms have had some big problems with that. But again, you know, going back to the social era, where if you can’t see what the product is, it’s you, there is an element, of course of the data that you are putting into GPT and various models is actually helping to support that. Now, good actors, like us, will be, you know, doing things like, I want to find out about products, ideas, and SEO, or whatever it might be. There’ll be plenty of people who won’t be doing that. So there is that but these are open models to a point. But where I guess that is going is, to private models and that’s already happening with API’s and because you are not going to put all of your company data into an open process where potentially could be hacked, this is not going to happen. So this has to be behind fire walls. So it’s a big security and enterprise level element to that for sure, but that’s already happening, you know, and because we’re seeing more and more integrations already over the GPT again, integrations or plugins, I should say – I have been on this last week. We desperately trying to  get into that and we have got some access and some beta access but that’s another game changer not to make things even more complicated because the problem we have got Jenny, the major issue is if the data is not understood and you do not know where it has come from it will hallucinate or, it might not necessarily be correct. Therefore you’re getting bad answers out, but that’s changing, because simply put, the integrations of the plugins. So Wolfram Alpha, for instance, was mentioned, which is searching for more scientific answers, correct scientific answers and is a plugin. So now you are not going to have all of these, right? Where’s it coming from? Is that correct? Is that hallucination?  Probably in three months, once they roll the beta out, tested it, got their correct partners in and moved on then that is going to go away. That is a  technology problem, that is not a business problem and that will be solved in my view.

 

Jenny Plant  30:46

You’ve actually answered the second question with this. And can you just repeat the name of that search engine for scientific data? Wolfram Alpha – the reason for that question was that with agencies who are listening to this and thinking, right, we are going to start experimenting or start using it. I suppose some are thinking how accurate is this going to be? Is are there citations that can I rely on what it’s telling me? So you’ve just answered that very well. Let’s move on to the whole thing about OpenAI opening their API to apps, websites, product services, everything. I know that people like Shopify and HubSpot are the early adopters aren’t they?. Just talk to me about what you are seeing, because this is also very, very recent, and how that’s going to change the game.

 

Roy Murphy  31:42

I think it changes everything. Because the minute you have an API, and that’s ingesting information, you can do anything. So that exactly means think back to the plugins, again, and to the API’s. So things like the I think the one that was mentioned in the demo last week, was Open Table. So again, more of that API ingestion, integration, and plugin is like okay,  getting back to the Google question, why would you go to Google for an answer? Even with the model attendings and the ads and everything, we get that, but you are not getting the contextual information. You are getting an answer, but then you have to ascertain is it a scam is a link?  Do I trust it? What is it? You’re in a box and that can be on an app, and this is to the API question, this can be an app anywhere this could be on  Open Table could be on the GPT website itself, which it won’t be in the future, by the way, people won’t be coming there, they’ll be doing other things and we will get onto that and that’s game changer. Because it’s basically the same as having Google with answers, in essence, and down the line real answers. So APIs. And I saw a couple of interesting thoughts on this during this week and last week, that OpenAI itself with API’s, plugins, and integrations is becoming a new type of business, which is a super aggregator. So it’s kind of what Facebook and Google and the GAFA, whatever the acronym is for the big four, or five. And they could really discontinue the whole lot. They could all go away, they probably won’t, but it could. Because if you’re using APIs, so you start with a search. So just say that the same paradigm exists and you start the search. But ultimately, you don’t have to go anywhere. And that search doesn’t have to start with GPT. You could be an Open Table, or you could be in Shopify, or you could be in Gucci. But it is ingested into it as an API. So they can go alright you can use it, you pay for it, of course. So you know, Jenny and Roy’s app has GPT and can do everything. But we have got to protect the nation and we have now got a hyper personalised focus and that’s what we’re doing. So that changes everything.

 

Jenny Plant  34:02

Wow, tell me a bit more about that. That sounds fascinating, this new business model, a new way of using it. So at the moment Roy, for example, you go on to OpenAI’s interface, you are using ChatGPT, or three or four, and you’re putting in your query, and it’s spitting out the answers. I mean, it’s just mind blowing. But in the future, like, let’s take the Open Table example. Obviously, they are plugging into the ChatGPT – API. And therefore?, tell me how that works like, what or where do you see us starting? Do you not see that we’re going to be starting a search on Google? How’s that? How’s that going to change? Just explain. Break it down for an absolute novice like me?

 

Roy Murphy  34:53

Yes. Like I said. I think we have got so used to – Google is a verb, right? We know that. You know exactly, it just is you know, the browser with a verb like Synthetic. (only joking!) All those things are verbs, right? So, but that’s because it is just usage. It’s just people have gotten used to it, I will  Google it. And I’m not saying that people are going to say that. –  I’ll GPT it. But I’ll tell you what, from the school playground, to business to bus and tube drivers to people in shops, the conversations I’ve heard, just listening and I’ve got kids, by the way, I wasn’t in the school playground, for no apparent reason, just to be clear. They’re all talking about it. That’s very unusual. So that’s the paradigm shift that I think is coming. Because we were just used to it and that’s why I think they are running, they are really, really terrified. Because this may seem like a small, because again, just to make the point, again, that we are talking about a Chatbot. All we are talking about the moment is a website, does it strip everything back, while it’s amazing is happening, but just talking about a box of little box on a site that you do something that was Google in 1996. And look what happened to Google, that whatever’s happened into their internal teams or their AI Deep Mind, they bought it 10 years ago, which they should have been on top of this and Facebook. They weren’t for whatever reason, I don’t know why maybe it is talent, maybe, Silicon Valley maybe its investment. So if you imagine Google didn’t exist, so it’s gone. There’s no more Google. Wow, okay, what do you do? Are you something else that you are being ? You can use another search engine. So that’s why I think this is fascinating. I think that we’re on the cusp of a massive change in how we interact with everything. And it started with the little Chatbot, which having been in the conversational AI game for several years, when we were talking about this stuff seven or eight years ago, and doing all the really hard work of natural language, language processing, and understanding and intensive occasions and trying to keep people on rails, and oh, my God, this would have been magic, absolutely magic. Now you can’t do everything there, you still going to have to have those, I’ve got a lot of powers in the conversational game. They are not going anywhere. But they’re very excited about this as well, and how this is going to get married together. But that I think is as big a shift as the Google search bar. And I am not saying it is going anywhere. But there is a propensity for the –  I think the phrase was the apex aggregator –  for OpenAI to be in the coming decade. And, you know,  we are all in on it. I’m not a maximalist in that sense and just like, it’s all that and nothing else. But I think as a business and agency, if you’re not really getting on top of this, and understanding what’s happening, I think you’re going to have a problem. I really do.

 

Jenny Plant  38:12

I totally agree with you. And this is just, I think, absolutely, it’s going to be ,I mean,  a bit of a cliche term, but paradigm shift in what we do. And for that reason, I’ve been banging on about it, particularly with the account management community, to say, keep ahead of this, keep an eye on this. So let’s bring it back to an account manager, and an agency owner, who may be listening to this thinking, okay, I get it, it’s going to change everything. At the moment, perhaps we’re not entering in that conversation with the clients. But clients, probably many clients are starting to ask about it. And I’d love to just kind of chat to you about initially, where are we going to see agencies having to adapt? and what’s your advice for them, so that they can start looking at testing new use cases, just in the agency business itself? Have you got any thoughts on that?

 

Roy Murphy  39:11

We were chatting earlier, actually, sorry. I just pinged over an account manager resource for you, just a very, very short one, before we chatted, which is saying this is a very basic 101 for account managers around generative. That took an hour. I didn’t use a developer. I have developed, but I am not developer. I ran a process of using ChatGPT , and Replit which is if you’ve heard, and I’m sure your listeners would have heard of, GitHub. It’s another version of GitHub which has, interestingly so, GitHub is also owned by Microsoft , so you can see why Google is scared.  So GitHub co-piloted AI, and it’s actually very well used and well recommended. It’s a good tool their AI part of GitHub. Their co-pilot Replit is a newer one, Google have just got into bed with them yesterday, so you can see what’s happening here. Because they are just trying to grab this information from different places, which is really interesting. But point being

AI is being democratised. This what is happening here,  if you take it just like a helicopter view, that’s what has actually happened, So for agencies, and marketing managers, and account managers, and so on, you have got to educate yourself. The basics are not difficult. You have to lean into this a little bit, and it can’t be the IT guy in the corner, the geeky guy or girl going, yeah, yeah, it’s a bit. Yeah, don’t really like it, No, you need to, because your clients are going to start asking for it.

Now, I will get back into that point. The other way of putting it is this is a talk about APIs. We are in an open environment business now. We are in partnership land and not just us. But this is where the future is going. It is not siloes it is partnerships. So you need to partner with us or someone else in this, whatever is appropriate for your business. But just talking on the macro level of Google and Replit and Co-pilot, and AI and all the super aggregators in the middle, agencies need to do the same thing. You are going to be toast, literally, and again, this is not against any content, or SEO agency or creative agency or development agency. But all the things we talked about touch on all those things, if I can put together a reasonably I got I’ve got very, very good guys. They will literally crying laughing at this point and be going, why do that, we’ll just do that better. The point wasn’t that, it was more, as a business owner, these tools are now available, even to demo and to prototype and go, Well, I’ve got an idea. Those are the things you should be taking to clients, who can’t even come talk to us or people like us, because this is where it’s going, that whole open partnership era, I think has started, I think this type of, kind of paradigm if you like, is going to accelerate it and that’s really the important point here, that you can’t ignore this. It is a tiny box in a thing, but it won’t be in three months, or six months, and a year and again, I mean that across anything, I mean that everything from PR agencies, to dev shops, to pure content to SEO to brand agencies to brand extension, doesn’t matter. If you’re not on top of this, you are going to have a problem.

 

Jenny Plant  42:45

So first of all, this resource that you created, and it took you an hour, which is unbelievable. First of all, thank you for doing it. It’s essentially for those that are listening, we are going to include the link, do you have the link handy Roy? Or should we put that afterwards?

 

Roy Murphy  43:00

We will stick in the notes, as I want to tweak it a little bit.

 

Jenny Plant  43:05

Okay, that’s fine. So basically, Roy has very kindly given us 101, for account managers for the terminology, and the acronyms that are used most frequently in generative AI so that you can sort of upskill yourself in terms of the types of terms that are used in this space. So thank you for doing that. Secondly, I think you have made a such a brilliant point about partnership, because I know lots of different agencies that are partnering with other agencies. And it makes total sense if your clients or you see that there’s an opportunity to be more forward thinking with your clients, partnering with an agency like Synthetic would help you quickly have that skill set. So Roy, is that something? because I know that you work with big enterprise level clients, you work with different kinds of marketing clients, but you also partner with agencies?

 

Roy Murphy  44:02

Yeah, absolutely. 100%? You have to, you absolutely have to, from a very similar business perspective, obviously it’s just a good idea because you can share resources, approach clients together and build teams out of whatever there might be on expertise and that’s agency land anyway, that generally happens. I think there’s been a shift that and I don’t know about what you’ve seen, but in the last three or four years that slightly “oh they are frenemies, I’m not going to talk to those agencies because they might”….  the scene to me has changed, I believe,. Maybe because of the area we are in, maybe because we have people often coming to us and saying we don’t understand this, can you help? Absolutely. That might be the case. But I think, there has been a bit of a shift in terms of understanding that whole openness around technology and society and behaviour change as well. Because if everyone’s talking about it well, why aren’t we doing it? because that’s a danger of being left behind as agencies because you need to be, you need to look at business models too, not just the way you work, but how are you co-creating with your clients? Because if you are a smart marketing manager and your client is going,” I’ve got five ideas and I’ve demoed them and got a prototype and what do I need you for again? What you do is great. And that’s brilliant. By the way, I figured out how to automate three of these things. So you know, that retainer, we’ve got kind of like, worried that it is actually a little bit expensive. and I’m going to halve it now.”   I’m playing the devil’s advocate but that’s the truth of it, right?

 

Jenny Plant  45:35

Listen, that’s the truth. I’ve just written a blog post and an email to my audience talking about this very thing that I think everyone needs to be armed, particularly because my audiences are account managers. Just about having the agency’s point of view on AI, because account managers particularly are having those conversations with their clients on a daily basis, they’re probably going to be the first point of call where this topic is going to come up. And second of all, I think I was trying to think about, you know, what’s the angle from the clients perspective? What is the account manager, going to be asked, first of all,? First of all, there’s a procurement website, where they give procurement people a list of questions, they should be asking their agencies about AI. So that’s the first thing that I can share the link to the questions and that resource. Second of all, I think you’re right, Roy, you know, we’re paying for this retainer, but we’re not seeing a reduction in the retainer, with you being a little bit more efficient with how you’re creating content for us. So that’s going to be a second conversation that I think clients are going to start having with account managers about, well, hang on a sec, this cost estimate isn’t changing. But I know, from my own experience, just using it myself, that we could be speeding up this process somehow. So if you’re not having the conversation, then either they think, I think they are going to be left with the thought that you’re not interested. You don’t know what’s going on. So I’m agreeing with you.

 

Roy Murphy  47:09

It’s agency 101, isn’t it? And again, you know, what comes up constantly with clients? that our agency partners aren’t proactive enough. And guess what? Now we can be proactive? Why do I need you so if you are not higher up the food chain, having those conversations, and understanding that you have got to be ahead of technology for your clients, you have to know more than they do. That’s what you’re for, your specialism doesn’t have to be an AI expert. But if you are an SEO, or whatever content agency or development shop, who is not going to come to you and say, have you used co-pilot? And are you doing that? And okay, even if it’s not, we want to double the work, we want more efficient work. So why is the cost the same? The output should be quicker now? Have I missed something? Or is the output quicker? Well, it’s not.. but these guys are saying it is. And they’ve obviously proved it and it makes them more competitive. And in fact, to be honest, I see that as a good thing, because it just washes out a whole bunch of, you know, not bad actors, but inefficient businesses, agencies too. Because, with a bit of Darwinism, you know, it’s evolve or die, isn’t it? It’s not the strongest, its the ones that are most adaptable to change, that is what happens. So this and  if this isn’t an adaptable to change shift, I don’t know what is. So keep it in mind Listeners,

 

Jenny Plant  48:29

I agree with you. I’m with you. 100%. So listen, I’m very conscious of your time, we’re coming up to the hour, which is just gone so quickly. I’ve just been riveted by this conversation. Any final pieces of advice, particularly for agencies tuning in? And also I’d love you to share contact details if there’s any agency that’s thinking, hey, I need to be ahead of this. I would love to have a chat with Roy, would you be open to having a conversation with another agency that perhaps is potentially looking for a partnership? And any kind of final pieces of advice?

 

Roy Murphy  49:01

Sure. Well, first of all, yes, absolutely. Come and talk to us always willing to have a conversation. It’s never a sales pitch, have a convo. Give us some pointers, partner up, whatever, we’re more than happy. We’re always open for that. Key things are educate yourself. And again, we mentioned it earlier,  when we were chatting, I’m not sure if we mentioned on the pod but this technology or the consumer facing part of this technology is now four months old. Now there’s a thing called the 10,000 hour rule, which you know about and for listeners, you can master anything for 10,000 hours, if you spent 40 hours a week on ChatGPT from the first of December, you’re already in the hundreds. So no one’s an expert, basically, but yes, there are people that understand it and they understand AI and how it works together and that’s obviously a skill, but you can get to grips pretty quickly and you should, so that’s the first place to start Chat to us,. chat to other people there are plenty of resources out there, literally go into Chat GPT and ask it, What are the best ….  okay right now the moment is not going to be, it’s not going to give you the up to date links to that. But that’s coming pretty soon, as I’ve already said, the integrations and the APIs, it’s already there basically just not launched yet. Two things to keep in mind, that I must mention,  GPT4 is seven months old. It’s not new. They’re already working on GPT5. it’s already underway. What’s out now was launched seven months ago, internally. So just keep that in mind everyone, that this is not the sharp end of the wedge by country mile. So that’s an important point to note. But do it ,do educate yourself. I would be remiss not to mention that we are helping for free to skill up social enterprises, charities, NGOs, and Nonprofits, we do a lot of work in that space. And we’ve got a initiative called AI April, we’re nearly full, because we’ve got one to one sessions only with those types of organisations to help them because we think that’s a really important thing. And I think it would really help them. So it’s any account managers or charities that are listening, who want or who knows someone that wants to get some upskill, from the experts on our side, we’re happy to help them. We’re running a whole bunch of workshops in April for that.

 

Jenny Plant  51:28

Amazing because I do know some agencies that are focused just on the charity space. So that will be particularly interesting for them, as well as some of these enterprises that are listening. That’s incredible. Anything else Roy? I know that I could spend another three hours talking to you about this stuff. But is there anything I haven’t asked you that I should have done?

 

Roy Murphy  51:51

No, you know, I’m loathe to recommend a bunch of tools we do from a commercial perspective, we do have our vetted database of, I think, nearly 500 tools that we use with our clients. So unfortunately, I can’t get that out. But we do work with it with our partners. But you can go and look for the relevant tools and so on. And I’m on LinkedIn, Roy Murphy, you’ll see my lovely AI face a much more handsome than in real life. Don’t be put off by it. But please come and see me, always willing to chat and to connect with people. My final thought is that its happening!. You know, it’s not going anywhere you need to get involved. Even if it’s just in your free time to bring back to your boss to say, I’m more efficient, I’ve done this.  You will probably get a pay rise. Right? Do that if nothing else, and I’m always willing to have a discussion and a chat, particularly any of the younger listeners as well. Always keen to give my tuppence when I’m not even asked for it and happy to mentor at points if people want a bit of career advice. So yeah, partnerships all over. Come and talk to us if you’re interested in generative and emerging tech. That’s what we do. It’s an exciting time. We’ll see you at work.

 

Jenny Plant  53:07

Thank you so much, Roy, this has been absolutely superb. We are going to put all the links in the show notes. I’ve been following Roy for a while and he shares some fantastic content. So make sure that you link in with him. Roy, thank you so much for your time.

 

Roy Murphy  53:23

Pleasure. Thanks.

Jenny

Author Jenny

More posts by Jenny

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.