In brief What makes a good AI ad isn't the speed of generation, it's the thinking behind it. We produced our first end-to-end AI-generated advert for re:Members, a B2B SaaS company, using Flora.ai as the production environment and a mix of image and video models. The work took weeks, not minutes, because the concept, the script and the creative direction all had to be right before a single frame was generated. If you're weighing up whether AI can make a genuinely good advert, the short answer is yes, but only with the same craft you'd bring to any other medium. Here's exactly how we made it, and what we'd tell you to watch out for.
What is AI brand visibility?
Humour us for a second. You type a prompt into a tool, hit generate, wait ninety seconds, and out comes a finished commercial. That’s the story the AI industry likes to tell, and it’s roughly the impression most people have. It’s also, mostly, nonsense.
We’ve just finished producing our first end-to-end AI-generated ad for re:Members, a company that provides Association Management Software (AMS) for non-profits, and fraternal organisations. The campaign we created, Escape the Same, highlights the opportunity membership organisations have when they switch providers. Surprise fees, clunky processes, disengaged members: most AMS platforms feel the same. It’s time for one that breaks the cycle.
The film we created for the campaign was built using Flora.ai – a model aggregator and creative playground that lets us swap between models as we go. So when someone claims that “AI made this ad”, the real question is, how much input did humans have in crafting AI to make this ad?
“AI raises the floor on what’s possible to create, but it doesn’t tell you what’s worth creating. That part is still on us.”
Kane Hawkins, Creative Director, Something Familiar
This is our answer to that, and to the bigger question we keep getting asked: what makes a good AI ad, rather than a forgettable one.
What What makes a good AI ad, and what makes a forgettable one
An AI-generated ad is a commercial whose core creative elements, the script, the visuals, the voice and the edit, are produced using generative AI models under human creative direction. The “under human creative direction” is important.
The wider industry is landing in the same place too as the IAB reports that the large majority of ad teams now use AI somewhere in the creative process, yet the research on results is more nuanced than the headlines suggest: human-led brand work still tends to win on recall and emotional engagement. AI is brilliant at volume. It’s still poor at meaning.
So the thing that separates a good AI ad from a generic one isn’t the model. It’s whether someone in the room knows what good looks like. Here’s what that came down to for us:
- A concept worth animating. We spent days on script and storyboard before we opened the tools – you can’t prompt your way to a genuinely good idea.
- Rationale behind every frame. Why this metaphor? What should each shot make the viewer feel? Where’s the emotional shift? Once that was clear, every generation had a job to do and consistancy to keep.
- Creative taste. Knowing when a generation is wrong, even when you can’t quite diagnose why, and having the judgement to throw it out.
- Direction, not just generation, on the voice. We treated AI voiceover like a real session: auditioning, directing, re-recording lines that didn’t carry the right energy.
- A destination that continues the idea. We built the campaign microsite to carry through the campaign – minimising disconnect for the audience and creating a seamless experience from ad to landing page.
AI will happily help you make something empty and audiences are already wise to stuff that feel machine-made. “Realness” is still the differentiator.
The real problem we were solving for our clients
re:Members is shaking up the world of association management software, the AMS platforms that professional associations, trade groups and alumni networks rely on. Their customers are stuck on legacy databases they hate. The systems are slow, ugly, and getting more and more out-dated.
But switching feels terrifying, because the database holds everything, decades of history. So our job wasn’t to sell features, or even the product. It was to make the pain of staying impossible to ignore.
That became Escape the Same. We landed on a metaphor closer to leaving a toxic relationship than upgrading software, off the back of a week-long concept sprint exploring four creative territories.
The vision was ambitious: alternative worlds, strange and beautiful effects, a dull grey office on one side and a bright meadow on the other. AI was the right tool for that, simply because it let us produce a film with the scale and strangeness of a much bigger budget.
How we made an AI-generated ad: the real process behind the thirty-second video
Beleive it or not, it wasn’t done in one take. Flora gives you remarkable control, but only if you know how to ask for it. We spent hours getting the lighting right on a single scene, iterating on how a character’s face moved, coaxing the model into a tone that matched the film.
There were dead ends. Sometimes a prompt produced a generation that was just wrong, and not for any reason we could pin down. Some scenes worked on their own but fell apart once we cut them together, so it turns out continuity issues aren’t only the remit of in-person shoots.
The voice took the same patience. AI voice tools have improved fast, but keeping cadence continuous is still hard. We auditioned multiple voices, directed them the way you’d direct any session, and re-recorded the lines that didn’t land. There’s an impression of effortlessness in the finished film. The reality was weeks of painstaking work.
We also built the Escape the Same campaign site to carry the click. Most paid campaigns send people to a homepage and lose them. We wanted somewhere the metaphor could continue, somewhere a visitor would land and feel the same emotional argument the ad started. The site is its own creative artefact, not just another landing page.
The tools and AI models we used to make the ad
As mentioned, we built the film using Flora.ai. Flora is a creative environment that brings multiple AI image and video models together in one place, so a team can generate, compare and combine outputs without hopping between separate tools.
The models that really worked for us were broadly Nano Banana, Seedance and Kling, across image and video generation.
That said, this space moves fast, so the specific versions we used then are already dated, new ones land most months, and the combination that gave us the best results then might not be the right pick now.
Can AI make a good advert? What this one proves
A creative agency using AI well can do things that simply weren’t viable before. A film at this level of finish, for a B2B SaaS launch, would have been impossible to justify two years ago. Now it’s the kind of work a mid-market client can afford.
But the work doesn’t make itself. The concept is still the concept. The script is still the script. The edit is still the edit. If you didn’t have a clear creative process before, AI won’t hand you one. It will just let you generate mediocrity faster.
We made Escape the Same to prove a few things, mostly to ourselves: that a UK creative agency can use AI as a real production tool, that the process is harder than people think, and that the results justify the time.
With this first ad done, we’re wondering whether the headlines around AI in ads should shift – to the fact that craft and care still matters, maybe more than it ever did. There’s a lot of slop, and we don’t want to contribute to that.
This thinking sits at the heart of our vision for AI, and it runs through how we approach AI creative production and briefing and brand visibility in AI search too.
Want an AI ad worth making?
Something Familiar is a Bristol-based creative agency and B Corp. We work with founders, growth-stage businesses and private equity portfolios on brand, digital and campaign work, and we use AI as a genuine production tool rather than a talking point.
If you’re weighing up what an AI-native production approach could do for your next launch, we’d love to talk.
Kane Hawkins is Creative Director at Something Familiar, a Bristol-based creative agency and B Corp.
FAQs
A good AI ad comes from the same place any good ad does: a clear concept, a reason behind every frame, and the creative taste to know when something isn’t working. The AI tools speed up production and lower the cost, but they don’t supply the idea or the judgement. The ads that land are directed by people who know what good looks like and use AI to reach it, not to replace it.
Not in ninety seconds. A genuine AI ad still starts with a script and a storyboard, then moves into generation inside a tool like Flora.ai, where you iterate on lighting, performance and tone shot by shot. Voiceover is auditioned and directed like a real session, scenes are edited together and checked for continuity, and the whole thing is shaped by human creative direction throughout. For us, the finished film took weeks of work, not minutes.
There’s no single price, because the cost sits in the thinking, not the generation. The AI tools themselves are affordable (for now at least…), but a good AI ad still needs concept development, scripting, art direction, editing and a place to send the traffic, and that’s where the work lives. As a rough guide, AI brings the cost of high-finish video well below what a traditional shoot would run, which is exactly why it opens up ambitious work for mid-market budgets that couldn’t have justified it before. The honest answer is that it depends on scope, so the better question is what you’re trying to achieve, and we’ll build the approach around that.
We produced our first end-to-end ad in Flora.ai, a creative environment that aggregates multiple AI models in one place. The models doing most of the work were broadly Nano Banana, Seedance and Kling, across image and video generation. The versions move fast and are already dated, so treat any specific list as a snapshot rather than a fixed recipe.
Yes, and it opens up work that wasn’t financially viable before, like a high-finish film for a B2B SaaS launch. The catch is that AI doesn’t replace a creative process, it amplifies whatever one you already have. An agency with strong concepting, scripting and direction can use AI to produce ambitious work affordably. An agency without those things will just make weak work faster.
