Selling Stories: Navigating the AI Production Transition with Ross MacRae and David Lyons

In the latest iteration of Selling Stories, Ross MacRae sits down with David Lyons to discuss what it really means to integrate AI into production at a practical level, not as an experiment or novelty, but as a fully formed creative workflow. Drawing on David’s background spanning executive production, photography representation, and cutting-edge AI production, the conversation explores how these new tools are changing client expectations, sales conversations, and the definition of craft itself.

With Mise-en-Scène at the forefront of AI-driven production, David brings a grounded, production-first perspective to the discussion, emphasising that while the tools may evolve, the fundamentals of storytelling, taste, and creative leadership remain constant. Together, David and Ross dig into how agencies, brands, and production companies are navigating this transition, why education has become the most valuable form of 'sales,' and where meaningful creative partnerships are headed next.

Skylar> Given that Mise-en-Scène has been hands-on building practical AI production workflows, not just experimenting, how do you position and 'sell' these services to agencies and brands? How does that conversation differ from how you describe more traditional production work?

David Lyons> The biggest shift is simply how many agencies now want to talk about AI. When we began exploring generative AI a little over three years ago, only the truly cutting-edge agencies were willing to engage in those conversations. I’d speak with senior leaders who were curious, and then bookings would come from very junior teams. But the mid-level agencies wouldn’t touch it.

That’s completely changed. Every agency, regardless of size, now wants to understand what AI can and can’t do. And that interest looks different in different regions. In Asia, clients honestly don’t care whether something is AI or not; they care about the budget and what’s possible. In Europe, many clients are already fully committed to using AI and prefer it over traditional production. In the US, it’s still early-stage: agencies want to test the waters, figure out the legalities, and experiment in small ways before committing.

Another shift is in how people find us. Previously our outreach through Agency Source / Bikini Lists used to get a 3–5% click-through rate; now it’s 27% on average. Agencies often start calls by saying, “We’ve been following your work for a while; we think it’s time we talk.” That didn’t happen three years ago. We also get a lot of requests from agencies asking if we can use AI to recreate celebrities or produce something simply because it avoids shooting. We turn those down. For us, AI is valuable when it enables ideas that would be too expensive, too complex, or simply impossible to shoot, not as a shortcut to avoid live action. That focus keeps the work more creatively interesting.

Ross Macrae> I don’t feel fully qualified to answer the technical side the way David can, but from my perspective, the biggest shift isn’t only about AI, it’s about brands coming directly to companies like Mise-en-Scène with creative inquiries. More brand teams are now in-house, and more agency talent is going brand-side, so the direct-to-production inquiries are growing.

That said, the ideas we’re seeing directly from brands can sometimes be more far-fetched. The type of work David and the team often decline because it won’t deliver what the brand thinks it will. Strangely, that’s good news for creatives. It reinforces the value of strong creative leadership on the agency side.

David> Yes, the direct inquiries we’ve turned down have almost all come from brands. Not the brands themselves, but the ideas presented. They just didn’t connect in a way that would deliver the right outcome. The exception is fashion, where in-house creatives are often exceptional, and that sector has always operated differently across photography and production.

Ross> When I looked at the Mise-en-Scène website a few months ago, what struck me was how hard it was to tell what was AI and what wasn’t. That’s a huge compliment, because it speaks directly to the level of craft.

And it reinforces something important: the conversation shouldn’t be about how the work is made. It should be about what is made, and why. When agencies try to do it themselves, it’s like the folks who were doing their own typesetting in the 1990s - it was technically possible, but the craft wasn’t there. Or like early computer graphics in the '80s. Just because the tool exists doesn’t mean the output is good.

From our side, we aren’t getting inquiries asking whether we list “AI-enabled production companies.” People come to us because they want the right production partner for the brief. They care about what the work is intended to do for the brand - not the tool used to make it. And I think that’s exactly what comes through in how Mise-en-Scène approaches their craft.

David> Exactly. At the end of the day, what matters is what’s produced, not how it’s produced.

Skylar> For Ross, along these lines, have potential Agency Source clients asked about lists for AI-enabled production companies, or similar? What tend to be the most-asked questions?

Ross> We don’t really get many inquiries asking whether we list AI-enabled production companies. And I think that’s largely because the majority of people who come to us are production companies themselves. Their questions aren’t about how the work is made, whether it’s AI, live action, or anything else, but about who the work is being made for.

Skylar> As the demand for AI-enabled production grows, you both have worked with agencies, brands directly, and other partners. For David, when you think about where your studio can deliver the most impact, how do you envision your ideal future client mix across agencies, brands, or other categories? Ross, the same goes for you, but regarding Agency Source’s impact on the same categories.

Ross> What we’re seeing on our side is that demand is shifting very clearly toward brands. Our brands-direct contact database in the US is now larger than our agency contact database, and that’s a change that has happened over the last five years. Of course, there are more corporations than agencies, but the growth still shows something important: production professionals increasingly want to work directly with brands, because that’s where the investment and opportunities are moving.

One major reason for that shift is the talent migration. A lot of ex-agency producers and creatives have moved brand-side, so the people doing the hiring already understand the production world, which makes those direct relationships easier and more productive. The briefs from brands may not always be as fully formed creatively, that’s something we talked about earlier, but the direction of travel is undeniable. More production people want to work directly, and that’s “where the money is,” to use the plain language of the industry.

For Agency Source, that means our impact is greatest in helping production companies find and connect with those brand-side producers, TV producers, and creative directors, many of whom previously worked at agencies in New York, LA, Chicago, and elsewhere. They already speak the same language as production partners, and they’re increasingly the ones holding the budget and decision-making power.

David> From our perspective, as demand grows, the place where we can make the most impact is with agencies and brands in the United States. As I mentioned earlier, Europe and Asia are already further ahead in adopting AI, while U.S. agencies and brands are still very much in the learning phase. That puts us in a strong position to help guide them.

A major part of our future impact will be educating senior leadership, showing them what’s possible with AI, what’s smart to use AI for, what isn’t, and what they should expect from the process. We want clients to understand AI as a creative partner, not just a tool that executes their pre-existing vision. That’s how you get the best, most interesting work.

Another area where we see meaningful growth is collaboration with existing production companies. If a production company has a director with a project that could benefit from AI, we can partner with them, not just as an effects provider, but in a deeper creative and technical capacity. The key, though, is that the director must truly understand how AI works and how the process differs. Without that understanding, the experience becomes frustrating for both sides. But when the alignment is there, the partnership can be incredibly powerful.

So looking ahead, we see ourselves working independently and collaboratively, with production companies, brands, and agencies, wherever the combination of the idea, the creative team, and the appetite for learning makes AI the right fit.

Skylar> Because both of your offerings combine creative production with deep technical fluency, it doesn’t always map cleanly to legacy models of sales or representation. How easy, or challenging, has it been to find people who can effectively communicate and sell your services in the market?

Ross> I’ll answer that one first. This is actually right at the heart of the 'selling stories' series we’re doing, because 'sales' seems such a dirty word in advertising and production, you can’t even really say it. Agency Source / BikiniLists has been around for 30 years, and over that time we’ve had many salespeople, reps, and account managers come and go. Because what we do is so specialist within an already specialised industry, we’ve always had to train our salespeople from scratch.

That means helping them understand how the advertising world works, how television works, how we work, how directors work, how photographers’ reps work, so our team can 'sell' our service correctly to our customers in what we call 'industry-savvy' terms.

And because we’ve got a lot of technology behind the scenes, people often think of what we do too simplistically - as if we “just sell a database of email addresses.”

One of our biggest challenges is convincing Production Companies and Reps that they are not just accessing email addresses, - the email addresses are there, but what our clients are actually buying is insight into who’s where, the job they are doing and what projects they’re working on. Getting customers to understand that can be a challenge.

David> That’s a very interesting and accurate point. When I use Agency Source, I focus on exactly that: information. The email addresses are a byproduct. What matters is the clarity around who works where, their background, and who might be open to understanding the kind of work we do. For us, for example, if we know AI can be particularly beneficial in the automotive industry, having that deeper context makes it much easier to identify the right people.

As for our own sales approach, it’s been challenging because it doesn’t resemble traditional sales models. My background is as an EP in production companies and in photography representation and production, so I know how those sales frameworks operate. While there’s some overlap, it’s not the same for what we’re doing now.

Our biggest and most successful form of 'sales,' if you want to call it that, is really education. We have to explain what we do, why we do it, what clients can expect, and how it might help them, and then step back. Traditional reps are used to pitching: “Here’s the director, here’s the portfolio, here’s why they’re right for this job.” That doesn’t apply here.

We know our work is of a high standard, but the sales conversation isn’t about the portfolio. It’s about explaining the usefulness and the process. That’s why a traditional sales rep wouldn’t work for us; they’d hear “an agency has an AI job” and run after it, but that’s not the right filter. We need people who can talk directly with leadership and help them understand the thought process and the workflow.

And truthfully, I get calls every day from someone wanting to ask “just one question” about AI, or our process, or what’s possible. That ongoing education is what eventually leads to work.

Skylar> With the rapid spread of AI creation tools, more companies feel they can experiment in-house. When you’re speaking with potential partners, do they assume they can build and manage these workflows themselves, or is there growing recognition that specialised talent is required?

David> More and more, we’re speaking with agencies that are setting up their own AI departments because there’s a directive from above that everyone needs to use and learn AI. Some people in agencies are overwhelmed and don’t want to take this on; others, usually creatives, dive in enthusiastically and produce something internally that looks good and is good for them. My view is: the more people using AI, the better. It helps the entire ecosystem.

But it goes back to the earlier analogy, just because an agency has an edit suite doesn’t make them great editors. For some things, in-house is fine, but when you want specialised, consistent quality, it makes sense to go to people who aren’t juggling other jobs and who do this full-time.

We work with AI platforms every day. We know the tools we used last Wednesday may not be performing the same today, so we switch. We’re constantly testing new platforms and revisiting ones we discarded earlier to see whether they’ve improved. That’s happening daily in the course of production work. When agencies build internal teams, we know they aren’t doing that level of continuous, technical evaluation, simply because we hear it from them.

Early on, agencies would ask us for our prompts, thinking that prompts were magic. But a single shot might go through an AI platform, then After Effects or Nuke, then back into AI with altered prompts; it’s a multi-step pipeline. Asking for the prompt is like asking a painter exactly which mix of paint they used. It doesn’t map to how the work is actually produced.

So we try to understand what they really want to know and present the information in a way that’s useful. I appreciate that these requests are often from a legal perspective.

We’re also beginning to see specialisation emerge within our own team. Some artists have mastered emotion in faces; others excel at environments. Some platforms we use require highly technical operators; others require artistic intuition. When we take on a project, we know exactly who should handle which part because that person will deliver it better and faster. If someone isn’t fast, that’s usually where they fail, we help them, but if the thought process isn’t there, we don’t have the time. It's hard to find and train the right crew.

AI is heading the same way CGI did: you have generalists, but on big projects, you need specialists, one person for liquids, another for hair, another for environments. We’re seeing AI move in that direction already.

And the biggest misunderstanding we encounter from clients is when they say, “We don’t want the AI to do that.” But the AI isn’t doing that, we are. We direct the AI at every step. Without human control, it doesn’t do anything meaningful. If they want to see what the AI produces on its own, we can show them, but they won't be happy. This reinforces that specialised talent is essential, not optional.

Skylar> For both of you, given your front-row perspective into both the buyer and builder sides of the industry, where do you see the most meaningful near-term opportunities as AI reshapes how agencies identify, assess, and engage production partners?

Ross> Very briefly, I think we’re still at quite an early stage. And I don’t believe AI, as an opportunity in itself, will be the determining factor in how agencies choose a production partner. It will continue to be the content that is produced, not how it was made, that decides how they identify, assess, and engage with partners. When they look at your website, they look at the work and make a decision based on the creative output. And that’s not to belittle the skill set or the ability to harness these tools, which you’re clearly doing at the bleeding edge, but in five years, I don’t think AI will even be part of the conversation around how a production partner is chosen.

David> I agree. AI will eventually become like asking, “What camera are we using?” It’s just a tool. But during this transition, there will still be filming involved, especially for brands that need to capture a location, a product, or a celebrity. AI will become something that enhances what’s already being captured.

One thing we’ve realised and are trying to communicate is that the same spend can produce a film that looks multi-million-dollar or a simple piece in a white room. It’s the same amount of work and the same budget. That gives creatives the opportunity to think, “I can actually do anything with this idea,” because within a certain range, the cost isn’t the limiting factor. In the ’90s and early 2000s, commercials had huge budgets and looked like it. Nobody can afford that anymore, but with AI, you actually can achieve that level of scale again. Younger creatives who never worked in that era will eventually realise that, and older creatives may pull out ideas they could never execute at the time.

In terms of how agencies will select production partners, I think they’ll continue to look at the lead creative, the director, and their storytelling, their taste level, how they compose a scene, and decisions around wardrobe and props. Those fundamentals won’t change. What will change is that the director will have to retrain part of their process to use AI and collaborate with the artists. The ones who do will be phenomenal filmmakers and commercial makers going forward.

I also don’t see AI taking work away from people. I see other jobs being created. You still need a stylist with great taste because we may scan the clothes they choose. You still need the makeup artist who can suggest how a face should look, because we may mimic that. None of that goes away. It’s very hopeful for the industry.

The problem is that people are running scared, and I wish they weren’t. I recently sent a marketing email and got a reply from an art director saying, “We don’t want any of your AI rubbish.” I wrote back with a long email explaining that we work with illustrators and retrain them with a new tool that lets them use their same taste level but work faster and more productively. It lets them take on more work and helps them futureproof themselves. It doesn’t eliminate their job unless they refuse to learn the new tool.

Change is the one constant in this industry. This isn’t the decimation of the industry; it’s another transition. Ross mentioned earlier that in the late ’80s, someone told him computers would never catch on in a studio. Today, it would be unthinkable to avoid using a Mac on a shoot. AI is the same kind of shift. Those who adapt will thrive.

For more information, visit Agency Source