In this episode of The Digital Edge, host Mark Reed-Edwards talks with brand strategist and copywriter Ben Afia to explore why AI is fundamentally a brand problem rather than just a productivity or automation tool. Drawing on decades of experience in brand tone of voice and his book, „The Human Business,“ Ben breaks down the dangers of relying on generic training data, which risks flattening unique brand personalities into a bland „regression to the mean.“
The conversation dives into a practical skills framework – Brief, Build, Back – designed to help marketers use AI as a collaborative thinking partner while retaining human judgment, taste, and empathy. From baking cultural standards directly into AI outputs to finding the hidden human angles in creative storytelling, this episode delivers an essential perspective for leadership on how bold creative work becomes infinitely more valuable in an automated marketplace.
Read the Full Transcript Here
Mark Reed-Edwards: This is the Digital Edge from Incubeta. I’m Mark Reed-Edwards. This podcast is about how you can balance technology and humanity. How as AI eats the world, you can integrate efficiency with empathy. We’ll talk with leaders from Incubeta and across the industry as we traverse the digital edge into tomorrow’s world. This time, Ben Afia is with me. His work centers on keeping brands distinctive with AI. He says AI is fundamentally a brand problem, not just a productivity tool, because generic training data can flatten differentiation and reduce trust if marketers outsource judgment and taste. Let’s get into the discussion.
Ben, welcome.
Ben Afia: Thanks for having me, Mark. It’s great to be here.
Mark Reed-Edwards: It’s wonderful having you here. And before we get into AI and branding, could you tell the audience a bit about your background and the path that led you to this current work?
Ben Afia: Well, I actually started my career in sales, but fairly quickly got into marketing. So I would say I’ve probably been in marketing for almost 30 years. And I landed at Boots the Chemist, where I specialized in brand tone of voice and managing agencies and writing and creative output. Got made redundant from that job, and that was 22 years ago now.
So I went freelance as a writer, specializing in brand tone of voice. And so since then, really, I’ve been working on creative copy, communications, and that really led into brand strategy. And it’s through that work really that I realized that the strengths of the brand, certainly in service brands, really come from the culture because it’s the culture that’s gotta deliver that brand.
And so this is when the kind of idea that I coalesced in my book that I published a couple of years ago, „The Human Business,“ where I started to really bring the idea together that if you really want to deliver the customer experience that customers will love, you’ve got to line your brand and your culture up.
Really, you’ve got to start from your culture to get the brand right, and that’s the only way you get the customer experience right. So that’s really where I got to, and I practiced that on some big, large-scale change programs over the years with companies like Vodafone, Aviva, Legal & General, Google a little bit, E.ON Energy.
So a range of mainly larger businesses, but companies where you’ve got hundreds, if not thousands of marketers, and certainly thousands of customer service people representing you and your voice to their market day in, day out.
But it’s really been the last couple of years that I’ve really then started delving into AI. Of course, it’s only been available to us for a few years.
Mark Reed-Edwards: Yeah, Yeah.
Ben Afia: And it was when I was heading up marketing for a renewable energy company and working with an AI specialist that I really started to get: there is really something interesting to this AI stuff from a brand perspective.
And the bit that’s really got me excited, is: you and I have worked over the years on brand guidelines, tone of voice guidelines. How do we describe the brand? How do we articulate the tone of voice so that people writing and speaking for us across the company can do that faithfully.
And, it’s very hard to implement that because, people have their own personality and we don’t want to turn them into clones. But the really exciting opportunity I see in AI is that you can bake your brand and your cultural standards into the AI so that the output is then naturally on brand.
And that’s just incredible, quite frankly. That’s a really exciting prospect. And so that’s really what I’ve started to focus on over the last year.
Mark Reed-Edwards: Yeah and your book came out a couple years ago and I remember you and I talking about it back then. Seems pretty relevant today- that intersection of brand, people, and AI. Were you anticipating that or did you see the early indications in AI that this kind of thing was going to happen, the intersection of brand, people, and AI?
Ben Afia: It would be nice to say that it was foresight, but it wasn’t. Making business more human has been really an itch that I’ve needed to scratch for decades. When I worked for other people, when I had proper jobs for employers, I always felt like I was a small cog in a large machine.
And I always felt that was quite frustrating, and I could see how that would reflect itself, that feeling would reflect itself with customers. That’s really where the human business idea came from, and that’s really what I’ve tried to do through language, through emotion, through behavior.
So really, it’s only fortuitous that actually I think that’s more relevant than ever when we think about AI. And the bit that I’m really concerned about is how do you remain yourself? How do you stay true to your brand, true to your personality? How do you respect the things that customers love about you when really AI is trained on generic content?
It’s trained on all of the writing that it has available to it on the web, on Reddit, stuff like that. The AI companies have trawled what is publicly available and some stuff that’s not.
So inevitably it’s gonna lead to a generic output. And so that’s the bit that I’m really interested in. How do we use AI so that it amplifies our personality, our values, rather than, make it bland?
Mark Reed-Edwards: You’ve said that AI is fundamentally a brand problem, not just a tech problem. So what are companies misunderstanding when they treat AI primarily as a productivity or automation initiative?
Ben Afia: I think the initial appeal of AI is efficiency, I think amongst marketing teams. And marketing teams are often some of the first in an organization to look at AI because they can see the opportunity, and marketers tend to be experimental, and they work with agencies who bring them ideas, so they just have more exposure and maybe a little bit more license to try stuff out.
And the initial instinct is to find efficiencies, where things are time-consuming. How can we write faster? How can we produce more? How can we automate process?
The bigger question for me though, ‚cause those questions aren’t really enough. The bigger question for me is how do we keep applying our own judgment, our own experience?
How do we maintain our voice? How do we retain our standards and our values?
If we move faster, does that help customers to trust us, or does it potentially reduce that trust? And my fear is that when every company, every marketer has got access to the same tools, trained on the same stuff, using the same vague prompts, the output can start to converge and start to sound similar.
And I see loads of posts actually on LinkedIn talking about, the last generic post that you read. It becomes increasingly difficult actually to distinguish decent, thoughtful, creative content from the generic. So I actually think efficiency is the wrong thing to focus on.
And actually, I’m working with a Fortune 500 client at the moment. They have 400 marketers or so, and we are building a series of AI agents that will help them all the way through the marketing workflow from end to end, everything from briefing right through to archiving creative. And what they’re looking for is automation, and I’m actually advising against it because
I actually think that a marketer needs to stay in control. Use the agents to speed up and improve the work that you’ve got at every stage, but you’ve got to still be applying your judgment. And my fear with marketing is actually that we lose creativity. You get a regression to the mean. If you end up with generic outputs, it’s not effective marketing.
Mark Reed-Edwards: Everybody’s zigging, nobody’s zagging, right?
One of the tensions I see in marketing is that AI rewards consistency. But great brands, look around, that list of companies you just went through that you’ve worked with, they depend on surprise, personality, and maybe even a bit of unpredictability. So how do you stop AI from flattening a brand into something generic?
And you were just touching on it.
Ben Afia: Yeah. First of all, I think it’s not just treating brand voice as a list of adjectives. It’s just not enough to distinguish. But the difference, I’ve been writing tone of voice guidelines for decades, literally.
Mark Reed-Edwards: And good at it.
Ben Afia: Thank you very much. And the challenge that we’ve always had in the past is that you can write some guidelines, but actually you’ve got to train people in ways to use them. If you’re asking people to write in a different way, for example, use active sentences rather than passive, that takes some time and effort and practice to apply. The difference with AI is that AI can do active and passive instantly. It’s no effort at all.
And so what I’m finding with defining tone of voice with AI is you can actually put a lot more of what you’d like to put in, a lot more of the color that you might have.
So I’m sure, you know, when you’re reading stuff, you spot some of the sentence construction that’s quite common or some of the individual words like quiet or quietly.
That one really annoys me. And that’s one of the AI tells for me. But, as individuals and as brands, we have particular words we would use and phrases that we would use, ways that we explain complex ideas that are particular to us that we’ve honed over time. How you sound when things have gone wrong, how- when you’re making an apology, how you handle tension.
So for me, it’s about training AI on your best material, and the lovely thing is that, in days gone by, I would have had to read 20, 30, 40, 50 pieces of communication to extract the tone of voice. But now AI can do that for me almost instantly.
Mark Reed-Edwards: Yeah.
Ben Afia: So if we gather the really strong examples of our writing, our stories, our founder stories, sales calls, our responses to complaints, examples of our values in action, then I think we can get the color and the nuance in our language.
And when I’m doing workshops in person with humans, and I use an approach called appreciative inquiry, or appreciative inquiry in the US, to hear some of these stories, that’s when we really get the unique language coming through. So the lovely thing with AI is we can build that into the AI so that we can amplify what makes us unique.
But I think you’re also asking about creativity, and I think that’s an interesting challenge. So if we rely on AI to write everything, how do we create anything genuinely unique? And I think it’s about workflow actually. It’s almost like many people are working still with AI as if it’s like Google Search.
Give it a question, expect an answer, bam, done.
Mark Reed-Edwards: Yeah.
Ben Afia: I look at it quite differently. I’m actually working on a skills framework that helps us to train people in a sequence of stages in learning how to work with AI as a thinking partner, because it’s much more than just giving it one prompt. It’s actually breaking a task down into a sequence of steps.
And so if we’re looking for more creativity, actually one thing that AI is really good at is generating creative ideas. So if you have a stage even before you start creating work where you’re defining your brief and generating ideas, you can actually use AI as a prompt partner to generate completely random creative routes. And I would do this in the past. I’d, let’s say I’m doing naming, for example. I might have six writers, and I’d send them all in different directions, let them work on it for a few days, and come back and see that diversity of thinking. We can do that much more quickly now with AI.
And you can tell it to go in different directions, and it’ll do that. So if you build the way that AI can do creativity into the process, you get much stronger output. And that, for me, is how you get more creative ideas. But the final part of that is not relying on it to create it for you.
So it’s part of the process, but the human still needs to be part of doing that work.
Mark Reed-Edwards: And it needs to be a chat, right? It’s not a search engine where you ask one question and you get the results, and you click on one and say, „Oh, that’s good enough.“ The chat is where it fine-tunes what you’re getting at.
Ben Afia: Yeah. Absolutely.
And it’s understanding your strengths, what you bring as a writer, as a marketer, as a leader, and what AI brings, and what different models, what different AI models bring. So how does Claude sound versus ChatGPT even when you’ve defined your tone of voice? For example, for me, Claude gives more nuance and more humanity in the tone of voice.
ChatGPT 5.5 has actually either caught up with Claude or Is starting to be. But they have different nuances and tone of voice, so learning to work with how they construct sentences, how they label things: Gemini puts stupid headings in whatever I ask it.
Really irritates me. So it’s learning how the different models work, but knowing what you bring and the strengths that you bring, understanding what AI can and can’t do, and then how you can work it to get the best results out of it.
Mark Reed-Edwards: And I think you’ve indicated it makes sense to use as many models as you can and see which one works best for you.
Ben Afia: Yeah. I’ve spent the last year experimenting with all of them. I do most of my work now with Claude. Claude is my friend, and I’ve created a whole range of agents that do all sorts of different work alongside me. I was gonna say for me, but I don’t really believe in trying to automate stuff just yet.
I think that’s too risky in terms of data and security and accidents waiting to happen. But I’ve worked across all of the models, and I’m constantly experimenting. And actually, one thing I am finding is it’s almost impossible to keep up with the changes, and as I’m sure you’re finding as everybody’s experiencing. AI’s moving so quickly, it’s very difficult to stay on top of it, and I don’t think we should really try.
So I’m actually now employing my son a day a week at the weekends. He’s in cybersecurity, but he’s helping me out at weekends in R&D, testing new models, running evaluations, helping me think through how I build my own personal tech stack and for my wider freelance team.
Mark Reed-Edwards: Fascinating times. Let’s talk about how marketing teams use AI. Many of them use it as a first draft engine. „I’ve got an idea. Can you flesh something out for me?“ But first drafts end up shaping the thinking down the road. So are we in danger of outsourcing not just execution, but judgment and taste?
Ben Afia: Yeah. Really interesting question. I think there is a risk of that. And so we need to be building habits that counteract it. And so I’m working on a skills framework that I call Brief Build Back.
And the Brief part is about the thinking that you need to do before you even start using AI. The Build part is then how you build in collaboration with AI, how you work with it. And the Back part comes after you’ve worked with AI, and it’s applying your own judgment and backing it up, whether you’re prepared to put your name to it.
The confusing thing about AI is that on the surface it’s really easy to use. It’s just chat. But actually learning to work with it is a lot harder.
So I’ll touch on the skills quickly. In the brief section, it’s about thinking how do we engage with AI and what’s the thinking that we need to do first? I’ve already talked about knowing your strengths and knowing what AI can and can’t do.
Knowing your strengths is knowing about what you personally bring. So for me, I’ve been working for 30 years. I’m a brand strategist, a copywriter, a project manager. I have a huge range of skills. So it’s thinking what strengths and skills do I bring that give me a particular perspective on this project that is actually gonna relate to the need that a client has brought to me?
And it’s then understanding how AI can help and where it shouldn’t, where I should stay in control or where I should do the work.
And then it’s about setting the vision. What does a good output look like? And what is the value of doing that work? Not just doing it for its own sake.
Who’s it for, and what do we want the person to think or feel or do as a result?
So setting that vision, building the context, so gathering the files. So there’s this the term context is a kind of an AI thing, but if you imagine a Google Drive or a OneDrive of files, the background that you would brief, if you’re briefing a creative agency, you’d need to give them some background.
So what context are you giving the AI? And then how are you breaking down a task so that you stay in control? How you draft the initial prompt, how you research or get AI to help you research, how you plan the project, how you review the outputs. So it’s thinking about how you make it part of your work.
I haven’t even got into prompting yet, and this is the thinking that we need to be doing, I think, before we even start using AI. And then the build part is about talking with AI like a person. I think of it like treating a creative, so working with a copywriter, working with a designer or an account director.
How would you talk to them?
And then prompting with purpose, so being really clear about what you’re asking them to do, how you want them to think about it, what format you want it to come back, and how you want them to behave here as well. Do you want them to push back? Do you want them to be direct, or do you want them to be encouraging and coaching?
Do you want them to challenge your thinking?
AI can’t read your mind, so you need to spell it out.
And then it’s breaking that work down, a large task down, so that it can work in sequence. So understanding how AI works and what makes it efficient, and then iterating with it, keeping going until it’s just right.
So I’ve been working on an outline for an agent for developing value propositions this afternoon. I got a first draft of an outline for the architecture yesterday. Looked at it this morning and thought that’s not quite right. The tone is a bit harsh. Doesn’t feel very human, doesn’t really feel like me.
I’m on version six this afternoon before it’s ready to send over to my client. And that brings us to the kind of the final, the kind of editor role, which is really you applying your judgment, checking the tone, checking the facts, checking it answers the brief. And, just very quickly, the final bit really is about owning that output.
Would you be prepared to put your name to it? Sometimes I describe it to people and say, would you be able to stand on stage and talk about it for two minutes?“
Mark Reed-Edwards: And also be informed about it because you get the feeling that a lot of… You mentioned LinkedIn earlier, and you get the feeling that some people don’t even read their posts before they copy and paste it into LinkedIn.
And I’ve seen, remnants of ChatGPT prompts in LinkedIn posts. That’s just shallow, isn’t it? Whether you have AI help you do your work or not isn’t the point. It’s whether or not it ends up being your work at the end of that.
Ben Afia: Exactly. It’s your name on it, and that’s a huge risk, isn’t it? And I’m paranoid. And it’s so easy just to cut and paste, and you leave a bit of the prompt or, „Is there anything I can do?“ Or, „Would you like me–
Mark Reed-Edwards: Would you like me to create an Instagram post from this? Yeah.
Ben Afia: It’s hilarious. So it really is about bringing what AI can’t, the creativity, the empathy, the stories.
So I’m delivering a speech tomorrow on my AI team to a group of business leaders and I’m talking about the Move 37 moment. So I start by talking about Move 37 and which just briefly, so in 2016 Google DeepMind took their AI called AlphaGo to South Korea to challenge Lee Sedol, at that time an 18-time world champion in Go, the ancient game that originated in China, has been played for thousands of years.
And it was thought to be a challenge that AI could not beat or probably wouldn’t for a decade or more at that time.
And so I start by talking about Lee Sedol, and it’s a very human story about man and machine. And Lee Sedol goes out for a cigarette at move 36 in the second game. And so I’ve got an image of his back to the camera, but with a cigarette in his fingers.
And I actually used AI. I have an AI speech writer that helps me prepare my speeches, helps me generate ideas, helps me write bits when I’m short of time.
But I actually draft them myself, and it had missed an idea. The AI had really not landed in the middle of the story, and you’ll know as a communications person that, if you can land people in the middle of the story, you engage them emotionally immediately.
So I actually can’t quite remember. I have rehearsed a couple of times so far today, but my opening is, „I want to introduce you to Lee Sedol. He’s an 18-time world champion Go player, and here he is smoking a cigarette.“
And my AI speech writer, which is trained on the way that I write speeches, had not seen that opportunity. It had helped me with ideas, but it had missed something that I just thought, „That’s a really human moment that I think my audience will relate to.“ So we still need to bring our creativity, our relationships, our stories, and know where it falls short.
Mark Reed-Edwards: The humanity, bringing the humanity in, which AI can nibble around the edges of, but the heart of it has to be you.
Ben Afia: There is one way that it can actually bring humanity in a way that is actually surprising. So I built an AI coach last year it was quite early last year, and I was working on a contract, and I was gonna be starting work on a website for this company.
And because it was a founder-led business, I knew that the tone of voice and the stories in this website would be really important and sensitive for the founders, because, it’s their story. And I was quite anxious about it. And so I knew that there’d be some tension for me, and I was wondering how I was gonna help my leaders to allow me to flex the tone of voice because they come from academic backgrounds, and so there’s a lot of academic language already.
And as my tone is very simple and very human and very straightforward and personal, so I was quite anxious about this, so I used my AI coach to coach me through the situation. Now, it didn’t actually give me the answer, but through the conversation, I was setting it up and saying: „Look, i’m anxious about how this is gonna go. I’m not sure how I should work it.“ I’m going on my previous assumptions. And one of the assumptions, one of the things that I got taught early in my career is that you should present your work face-to-face. You should talk things through with people.
And so my instinct was that I would come up with a first outline of copy for this website, and I should sit down with my sponsor, with my client, and talk it through with them face-to-face.
But through this conversation, the AI actually helped me to see, what if there was a different way to look at it?
What if my client is actually an introverted thinker and a reflective person? And what if that person might be intimidated by having to go through this face-to-face? And what if they need time to reflect before they respond?
And I thought, do you know what? I’m not gonna book a meeting. I’m gonna send it over.
So I sent this first draft and I got feedback almost immediately, and she loved it. She made some small amends, but she accepted the tone that I’d given, and it worked so much better than I ever expected, and AI had helped me to see that other side.
Now, a coach would have helped me to do that, but I didn’t have a coach available at the time. I had AI.
So for me, that’s actually a very human application that AI has brought me to, a human insight that AI helped me to get to. So if we use AI to see different angles and see things from our audience’s perspective or from another person’s point of view, I think it can be really powerful.
Mark Reed-Edwards: So, you know, I read about AI every day, and it seems like there’s kind of a tension between companies that say “ we’re all in on AI,“ and then you read a bit deeper and you find out they’re all stuck in pilot mode. So is that mostly a technology issue, or is there a cultural issue around fear and uncertainty inside companies?
Ben Afia: I do think there’s a lot of fear, and I genuinely don’t think it’s the technology. The technology is actually relatively straightforward. AI itself is fairly easy to use. Building a knowledge layer is a little bit more technically complex.
But in terms of day-to-day use for individual people, I think it’s about humans and their relationship with it.
And their relationship with their role and their relationship with the organization. And I think a lot of organizations are piling in to pilots and they’re not stepping back and going, „What are we actually intending to do here? And what does it mean for the people in the organization?“ Because I’m sure every employee under the sun is thinking, „What does AI mean for my job?“
And that is a very valid fear.
Because CEOs and boards are under pressure to deliver quarterly results and to deliver efficiency in order to improve profitability. So it’s easy to see how short-sighted leadership could say, „Great, we can get loads of automation. What if we get all of our teams to train our AI in the way we do things, and then we can get rid of them?“
And even if that’s not the intention, I think people are probably wondering if that’s the case.
So I do encourage organizations to be really purposeful and come up with some sort of statement of intent around AI, a policy about what the intention is.
But also then to introduce people to AI in a way that really works.
And I organize a monthly drinks with friends on my street. It happens to be tonight. And I can remember, ever since the beginning of this year, we’ve been for our monthly drinks, and I’ve got one friend who works at a global law firm.
I’ve got a friend who works at a global manufacturer.
There’s people in all sorts of businesses, large and small. They’ve mostly been given Microsoft Copilot licenses, and they’ve mostly either not been given any training or been given really basic training „Here’s how you write a prompt.“ So I think the thing that’s missing is helping people to understand how to think with AI rather than just use it like Google Search.
Mark Reed-Edwards: Put yourself in 1981 and someone gives you a personal computer and says, „Here, use this.“
Ben Afia: Totally. Funnily enough, I think it was around 1981 that my grandad bought a Sinclair ZX Spectrum, I think, one of the very early computers, and said, “ Go on then, Ben. Go and work out how to use this so you can tell me.“ So I took it away, and all I did was play Frogger.
Mark Reed-Edwards: Exactly.
Ben Afia: So exactly that situation, and this is revealing my age. But yeah.
Mark Reed-Edwards: I’ve got one last question, and I wanna dig a bit into your book, „The Human Business.“ In it, you emphasize the human side of business and change, and we’ve talked a lot about that. AI systems are trained on what already exists, and maybe that then reinforces convention, conventional wisdom.
But memorable brands often come from doing something unexpected or even uncomfortable. So in an AI-driven world, does bold creative work become more valuable or harder to justify internally? What do you think?
Ben Afia: I think it becomes infinitely more valuable. My fear is marketing teams all over the globe start using AI to deliver more faster, and because it’s easier on the surface, they’re applying less judgment and they’re bringing less of themselves to it.
And I suppose even amongst, trained marketing teams, I find that understanding of brand and differentiation can be not as strong as it might be.
So people that really deeply understand brand instinctively understand that the point of marketing is to stand out in the minds of your customer.
And so bold creative work is essential to creating that.
I think we’re gonna have a lot more generic marketing, and that for sure is gonna increase the value of bold creative work.
We need to be focusing on how we stand out.
I don’t think it’s harder to justify at all. I think it’s vital. It’s gonna become increasingly crucial. And in fact, I’m seeing brands now popping up who are leaning on their humanity and saying, „This is human created.“ There’s a podcast I listen to when I wake up in the middle of the night, and the description now says „Scripted by humans, read by humans, human produced,“ something like that.
Mark Reed-Edwards: Yeah.
Ben Afia: I think we’re gonna see more of this, and this is where, brands like Etsy come in and where they started. We have an aversion to things that are machine-made, and we wanna know that there’s some soul to something.
There’s some interesting research. The Nudge Podcast last year, maybe October, November, touched on some research where creative work, writing, art, and music was tested out with a varied audience. And what was really interesting is that when they didn’t know it was created by AI, people actually preferred the AI results until they were told it was AI, and then they didn’t like it.
Now, I think that’s something that will pass over time.
We’ll just get used to seeing things that have been created by AI, but as marketers and creatives, we need to be really careful that we’re not slipping into producing stuff that people just go, „That’s clearly AI. I’m not interested.“
If we’re not differentiated, if we don’t stand out in our customer’s mind for some reason, we don’t have a brand and people are just gonna stop buying us, or we become price-led.
Mark Reed-Edwards: Ben, when we discussed having you on the podcast, I knew it would be a really interesting discussion, and this exceeded my expectations. Thank you so much for joining me here on The Digital Edge.
Ben Afia: It’s been my absolute pleasure.
Mark Reed-Edwards: A great discussion with Ben. Let’s heed that warning at the end. Sure, use AI, but let’s not slip into producing stuff that is obviously AI. Keep the human element. Thanks for being here today. I’m Mark Reed-Edwards. Join me on the Digital Edge next time.
Speakers: Host: Mark Reed-Edwards; Guest: Ben Afia
Take A Deeper Dive Into The World Of The Marketer
We are entering an era where the rules of marketing are being quietly rewritten – not by tools, but by how decisions are made. In Incubeta’s latest research report “The Marketer’s Confidence Paradox” – conducted among marketing leaders in Retail and eCommerce – it’s revealed that while 70.4% of marketing leaders are confident their budgets are deployed effectively, 41.6% acknowledge a significant portion of that investment fails to deliver its full value. Incubeta’s research explores the gap between perceived performance and real growth, providing a roadmap to bridge the divide.
