In this episode of The Digital Edge, host Mark Reed-Edwards chats with Tom Williams, Global Creative Director at Incubeta, to explore the evolution of creative in the age of AI. Moving beyond the „prompt engineering“ hype, Tom explains why traditional art direction and storytelling remain the essential anchors for high-quality brand creative. From leveraging tools like Veo 3 and Seedance 2 to maintaining brand governance through custom AI frameworks, this conversation explores how to use technology as a creative partner to scale imagination without sacrificing the „human-centric“ narrative that drives true consumer trust.
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 to leaders from Incubeta and across the industry as we traverse the digital edge into tomorrow’s world.
On today’s episode, we’re joined by Tom Williams, Global Creative Director at Incubeta. He says nothing makes him happier than creating vibrant pieces of art and design to help brands engage with their audience across various platforms and media. Together, we continue our look at AI – this time with a focus on creative from a human perspective.
Tom, welcome to The Digital Edge.
Tom Williams: Thank you for having me.
Mark Reed-Edwards: Can you share a bit about your creative journey and how you came to Incubeta?
Tom Williams: Yeah, certainly.
I wouldn’t say I’ve necessarily had the most traditional journey into the field. Creative has always been something that I’ve enjoyed and loved since a young age. I’ve always enjoyed painting and art – something I excelled in at school and college – and I started off in that traditional route.
I came more from the art side of creatives, so graphic design and art direction. I did have the misfortune of graduating at possibly the worst time, in 2008, just as we were in the economic downturn. But as luck would have it, I did end up starting my career as a junior graphic designer working in-house at a charity many moons ago.
I’m old enough, and of the generation of older millennials, who still remembers the more analog version of design. Then I made the inevitable switch over toward more digital, social-led creative campaigns, which I think has been quite interesting – living through that change in the industry – because I think we’re now on the cusp of undergoing a new revolution as AI systems start to become more and more integrated into our ways of living and working.
It’s actually equipped me to eagerly take on the challenges of having to adapt to a new way of working. Understandably, I think there are reasons for trepidation, but I’m excited about what’s around the corner.
Mark Reed-Edwards: Yeah. The old cliché is there’s one constant – change. And we are in a business that thrives on change.
Tom Williams: I think knowing that there is always that need to adapt – if you rest on your laurels, you get left behind or you become irrelevant. So I think stand up for what you believe in, obviously, and don’t fall for everything.
But I think there needs to be the anticipation that the industry is always changing and evolving, and you have to go along with that at the same time.
Mark Reed-Edwards: You mentioned the magic buzzword AI. We’ve talked a little bit about creative on this new season of The Digital Edge and how creative works with data. And I guess this is a variant of that question:
How does creative work differently in the AI age, and then how do you set yourself up to have a digital edge with that creative?
Tom Williams: Yeah, it has been fascinating – I’d say three or four years now.
I remember when one of my designers first came to one of our inspiration sessions and was like, “Oh guys, there’s this new AI tool and we can use it to produce images.” We were like, “Oh, okay. That’s interesting,” and fired up Discord and Midjourney.
He showed us these somewhat abstract, avant-garde interpretations of what was AI art back then. And I think we found it interesting, but there had been these long murmurings of AI and its development. I think everyone saw creativity as one of the last bastions of human inspiration and interpretation – and a machine could never replicate that.
But obviously, as it has evolved and developed at this real exponential rate, I think a lot of people have started to see the benefits and the pitfalls of leveraging it for creative. As I say, there is the good and the bad.
I think for us, the speed at which it’s allowed us to actually build out and visualize concepts has been a real game changer. Before, it used to take us days or weeks to work on a key visual or start putting together a storyboard for a film. And now that can be done in hours or a day at most.
But what I think it also highlights is that there is still the need for expertise and specialization and experience, because it is that experience and knowledge that we actually build upon and use to guide AI – not just reference, but actually provide art direction.
So it’s really become, for us, a bit of a creative partner, which has allowed us to leverage scale and speed like we never did before, but not at the expense of craft and storytelling and art direction, which I think is still essential to what we do.
I just think it’s an incredibly powerful tool, and I think we’re lucky enough to be at a crossroads in history where, for the first time ever, the limitations are not the tools or the budgets or the resources or the time, but actually your imagination.
But it still depends on your ability to adequately communicate and verbalize instructions and provide exact art references and images. And that’s how we’ve really managed to maintain quality and consistency while starting to integrate AI into what we do.
So yeah, it’s been a really exciting time – something the team has really embraced. Again, I think everyone’s been encouraged to learn and adapt and make the most of it, and use it with intent – not simply just because it is the buzzword. It is done with purpose, not at the expense of quality.
Mark Reed-Edwards: Do you find yourself part creative director, part prompt engineer when you’re thinking about engaging with one of these platforms?
Tom Williams: I find the term prompt engineer a strange one, especially in the realm of creativity. I find it just a label that people have slapped onto the process.
For us, I think it is still art direction. I approach it like I do art direction. I’m using the same descriptive and technical language to describe the composition of a scene and lighting as I would if I were trying to brief a photographer or a cinematographer or indeed my designers.
And I think obviously LLMs are built off that learning and understanding. So as I said before, I think it comes down to how concisely you can express and communicate your vision into that prompt.
That’s not to say there isn’t some technical understanding. I do appreciate AI is not a person, and so you do need to speak to it in a slightly different way to get more consistent outcomes aligned with your way of thinking.
Through a lot of trial and error, we’ve certainly learned how to start doing that – how to take a more formulaic approach, building context and structure into our prompts.
As I say, I think the biggest struggle for a while has been consistency, but learning how to reuse terms and phrases – and the outcomes those generate – and use that for future prompting has certainly helped in that field.
The last three or four years have been a lot of trial and testing, and I think we’ve gotten to a place now where it’s a viable product for our clients. But what we’re trying to do now is actually build a more comprehensive framework and systems where we’re not having to constantly re-prompt, but instead have a more modular approach.
This allows us to retain brand governance, avoid model drift with characters, and make sure scenes and art styles are more consistent from one to the next.
In part, as the platforms themselves evolve and cater for that too, it means we’re not having to be as detailed in some of our prompts as we maybe had to before. It’s starting to do a bit more of that legwork, but we are still the ones in control at the end of the day, and we still need to provide that creative direction.
It is like anything – you get out what you put in. And it’s not a magic wand that you simply wave and expect to get Hollywood-like results. You do still need that craft and those experts with that background knowledge to provide that context and get that output.
AI is definitely a multiplier, but it’s certainly not a shortcut. And I think brands who take that lazy, generic, and somewhat uninspiring approach are going to be the ones that fade into irrelevance.
I think at the heart of it still, while we may be using machines, we are actually still creating a human-centric narrative and storytelling, especially in advertising. While the way in which we go about the job has changed, the purpose and what we are doing hasn’t.
We’re still trying to build a connection between brands and their audiences, maintain relevance, and make sure that people recognize the people in our ads as themselves to encourage interaction and ultimately performance and purchase, which I think it is enabling us to do.
Mark Reed-Edwards: This kind of underscores the importance of the brief, right? The creative brief that you use as your guide when you’re doing a project for a client, and a brief that’s not complete will lead to incomplete results. And I guess it’s the same when you brief your AI tool.
Tom Williams: Absolutely. You get out what you put in, and I think making sure you are concise and very direct in the context that you provide. Much like when you receive a creative brief, the more information you have to hand, the more references you have to act as a benchmark, the more that’s going to guide and direct you to the results you want.
If you’ve got something that’s very hazy, it could be interpreted in a number of ways. That’s not helpful to anyone. It’s not helpful to the person providing the prompts. It’s not helpful to the LLMs or the models that we are using in terms of the output that they’re then going to generate.
It’s like we say with all clients: the more you put in, the more you get out. And it’s certainly the case with AI as well.
Mark Reed-Edwards: So of those new technologies and apps, which ones excite you?
Tom Williams: We’re quite fortunate that we’re a longstanding Google partner, so they did give us a lot of early access to things like Nano Banana and Veo 3, which at the time were massive game changers just in terms of, as I say, quality.
Veo 3’s ability to lip-sync audio and do dubbing with video – which I know we take for granted in traditional filmmaking – but was something that hadn’t really been seen before in video generation, really set the benchmark for future models.
Fast forward six months – not a lot of time in the world, but in AI that’s like a lifetime – I’m very excited by some of the Chinese models coming out. Seedance 2 in particular again seems to have just pushed the needle that little bit further.
In terms of consistency from frame to frame, the camera controls that it now offers are really pushing this sense of realism – the kind of things you saw in more traditional animation and filmmaking – which I think is why Hollywood was a little bit scared and knee-jerked with a lot of cease and desist, understandably.
But I think the progression it shows, and if it truly does live up to the hype that it’s already generated in the community, it’s really going to allow us to focus, as I said, on that storytelling element and not have to worry about model drift or inconsistencies in characters from scene to scene.
It’s going to do that heavy lifting for us, and again we can actually focus on the visual and narrative that we are trying to drive, but do things that before would have cost millions in special effects.
Again, now with the right imagination and prompting, you could do it in your bedroom in half a day and create some really cool stuff.
Which ultimately, I think, is where we’ve seen benefit – where you are limited on resource, time, or budget – for all the things that afflict marketers since the dawn of time. I think this has given us a viable option in leveling the playing field in certain respects.
Other things that I’m quite excited to see coming in: again, I think a lot of people have gotten used to this sort of prompt-and-repeat cycle with many of the platforms, but we’re quite excited to be really developing these responsive systems.
And the way we’ve been doing that is using something called LoRAs, or low-rank adaptation models. And that’s actually been really exciting for us and the brands we’ve engaged with because it allows us to take a more modular approach.
We can almost plug in their own brand guidelines, their governance, their tone of voice, and again that influences every prompt across the lifecycle and the workflow without the need to start from scratch again and again.
It also means we can take a more collaborative approach. Different team members can plug in at different points of that workflow and we’re still going to get those consistent results, which I think is super important for our clients. Especially where, again, I think we should be using AI to build things that are still truthful, that are still true to life and still adhere to physics.
Obviously, we can play around with special effects, but audiences don’t like being lied to. And I think for a lot of brands, it’s important that their product is represented like for like. Again, we’re not trying to pull the wool over a consumer’s eyes, and we maintain that trust – and I think we do that through realism.
So again, seeing the ways in which the models have evolved and are doing a better job in terms of motion and physics is something that’s really exciting.
Mark Reed-Edwards: So that someone watching can say, „That really could have happened.“
Tom Williams: Yeah, I think look at some of the fashion brands we’ve worked with, or even automotive – and anything that has a physical presence, like a car, a shoe, or a handbag. Advertising is about selling the best version of the truth, but you’re not lying to a consumer.
People want to know that what they see is what they are going to get, and I think we can still achieve that with AI. We just have to avoid getting lazy and cutting corners.
And again, what’s quite exciting about the new models is the amount of references you can provide them with – both in terms of video, imagery, and audio – means you can be almost one-to-one exact replications now from every angle of that product, which I think is super important and something that was a challenge before.
You had to jump through a number of hoops and a couple of tricks to get around it. But again, I think as the models are learning, they’re doing a much better job with that now.
It still needs to put the consumer at the heart of the story and clearly demonstrate the tensions that they are feeling, or the relevance to their lives as to why this product or service is applicable to them.
And again, I think we only do that by portraying the truth.
I think the moment we don’t is the moment that we lose that trust and it all falls apart.
Mark Reed-Edwards: So how do you, in this kind of brave new world, balance output and performance, and how do you keep the human touch with AI so omnipresent?
Tom Williams: Yeah, it’s a bit of an oxymoron or a juxtaposition, if you will – how do we sound human when we’re using a machine at the heart of our generations?
And again, I think it’s not losing sight of the fact that we are using AI as a tool. It’s the tool, not the product.
I think more now than ever it’s a necessity that we actually listen to and respond to our audiences in real time and adapt to their desires and behaviors, rather than simply pushing what we believe is right or what they’re going to engage with.
And you mentioned obviously the impact that AI has had on data. In turn, that feeds into the creative itself and allows us to react in real time to maintain that relevance with our audience and do it at a scale we’ve never been able to do before.
But at the same time, at that scale, how do we still keep it feeling personal and resonating with the individual rather than the masses, even though that is obviously who we are at times trying to market to?
I think it comes back down to human insight in terms of culture and nuances, and using AI to still tell those stories of connection and build that relationship between brands and their consumers, rather than distract them with a frivolous use of AI.
I think that’s where realism comes back in as this sort of engine of trust and why we have to maintain that through all of our generations and maintain that level of quality and expectation.
The last thing I want is AI to be a race to the bottom. I think it’s why the importance of those experts and those practitioners in craft still need to be at the reins.
I think, plus, it’s good that it has leveled the playing field. In doing so, it might have put the tools in the hands of those who are maybe not quite so capable of wielding it.
And I think as long as we don’t accept that as the new standard, we can keep those levels of quality high, and in turn that is going to drive performance.
We don’t want to simply become a factory that just churns. I think it’s specifically about listening to the data and interpreting how we actually turn that into engaging content.
Mark Reed-Edwards: It’s an interesting point of view, Tom. I think I’ve learned a lot about what you do in your day job. Now I’d like to learn a bit more about Tom Williams the person.
The last couple of questions are aimed at that. Can you give me an example of when you failed and how it impacted your life?
Tom Williams: Yeah, so not necessarily a failure, but I think because of the somewhat unorthodox entry and route that I’ve taken into the field – which has led, I guess, to my current standing and position – I think for the longest time that left me battling this feeling of inadequacy.
And as I know a lot of people deal with this almost imposter syndrome – “Oh, should I have done this? Should I have studied this at college or should I have taken this job?” – and so I think that kind of self-doubt, for quite a while, I had to try and overcome.
But in turn, I don’t call it a superpower necessarily, but that has driven me continuously to constantly learn, to constantly want to improve myself, both technically and practically.
Whereas maybe, if I hadn’t done that, would I have had a level of complacency?
And I think that initial maybe stalling of my career in the early years has actually been that driving force now throughout my career in terms of my work ethic – always wanting to stay on top and improve and learn.
And that willingness to adapt to the constant of change in order to, as I say, make sure I’m as much as I can be ahead of the curve, and always encouraging and empowering others.
I guess, again, I think maybe not having the best start to my career, I’m very grateful for any opportunities that I was then given. And I think in turn I want to pass that on to the next generation of creatives and give them a chance and empower and encourage them.
Because yeah, it’s not easy out there. I think showing that kindness to one another is only going to be best for the industry and continue to keep those standards high.
Mark Reed-Edwards: So last question, related to what we were just talking about, I think – if you could share one piece of hard-earned wisdom you’ve learned over your career, what would it be?
Tom Williams: I don’t know if I’m full of too much wisdom, but I’ll give it a go.
It’s not an easy world that we live in today and there’s a lot of hardships, but one thing I have learned is that skills, tools, technologies, theories can be learned.
I think it’s not necessarily an inherent talent, but I think what’s harder to teach is that dedication and that commitment to want to drive and push yourself.
And certainly as I see the next generation coming up, that’s what sets the people who succeed apart.
I think you see plenty of people who’ve got that raw talent, but maybe not that same drive. Unfortunately, that’s not enough to get you ahead.
You have to be willing to adapt and flexible enough and humble enough to know when you actually need to improve and find ways and seek out help and collaboration in order to achieve that.
But yeah, I certainly think as long as you work hard and you’re dedicated and you help those around you and take help when it’s offered – regardless of what comes your way – it’ll always work out.
Maybe not for the best, maybe not for the worst, but I think with that dedication and that work ethic, you’ll certainly win more than you lose.
Keep going. Don’t give up. Follow your dreams and trust your gut ultimately.
Mark Reed-Edwards: It’s a great way to wrap up this episode. Tom, thanks for joining me on The Digital Edge.
Tom Williams: Thank you very much for having me. It’s been a pleasure.
Mark Reed-Edwards: Is there anything better than a discussion with a creative director, especially one of Tom’s caliber? The ideas flow, the solutions bubble up, and the clients are happy.
I was really intrigued by the way Tom is using AI with human inspiration.
Thanks for being here today. I’m Mark Reed Edwards. Join me on The Digital Edge next time.
Speakers: Host: Mark Reed-Edwards; Guest: Tom Williams