The operational reality of marketing inside an AI company

By Tina Sang, Marketing & Growth Lead at fal

I spent last week at Cannes Lions International Festival of Creativity, where nearly every conversation eventually landed on AI.

Usually the same questions came up:

Will AI replace creatives? Will everyone become a filmmaker? What happens to originality?

Martin Sorrell, founder of advertising giant WPP, argued that the industry may be focused on the wrong AI debate. I agree, but for different reasons. What bothered me was how detached the conversation felt from the reality of using AI inside a company.

I lead marketing and growth at fal, a generative media infrastructure company. That means I spend most of my time sitting between frontier models, product launches, customers, and go-to-market.

One thing working in this space has made clear to me is how artificial the distinction between marketing and growth really is. In many companies, marketing is still thought of as awareness and brand, while growth sits downstream in signups, activation, and retention. That separation feels increasingly outdated.

In our world, a product launch, a customer case study, a viral creative campaign, or a technical tutorial can all drive the same thing: product adoption. The question isn’t just whether something gets attention, it’s whether that attention compounds into usage.

That lens also shapes how I think about AI. Once marketing and growth are treated as one system, the highest-leverage work stops being content production and becomes judgment: what to prioritize, what to say, when to say it, and where it will compound. That’s why working in this space has made me more skeptical of automation, not less.

People assume that because we’re an AI company, our workflows must be heavily automated. Agents for everything, autonomous content pipelines, and a constant stream of generated assets. But that’s not really how we work.

One thing I wish more marketing leaders understood is that adopting AI doesn’t have to mean handing over your workflow to autonomous systems. In practice, the most useful applications are often much narrower, more practical, and far more controllable.

At fal, we use AI selectively at points in the workflow where speed matters and the downside is contained.

Usually that means:

• production mockups
• creative exploration
• fleshing out rough first drafts
• quickly visualizing an idea so the team can react to it

That last one has been especially useful. The value, at least for us, isn’t usually the final generated asset. It’s how quickly AI helps turn an instinct into something concrete.

A lot of ideas die because they stay verbal. A few weeks ago, someone dropped into our Slack: “What if we did custom FIFA World Cup player cards?”

Before AI, that might have become another bullet point sitting in a backlog. Maybe it would’ve done well, maybe not. It’s hard to tell when an idea only exists as words.

Now we can mock it up in twenty minutes. The conversation immediately changes. People stop reacting to an abstract idea and start reacting to something tangible, which dramatically speeds up decision-making.

One of AI’s most practical contributions to marketing is reducing the cost of moving from instinct to artifact.

Our team spends a lot of time on X, partly because the AI ecosystem lives there, but also because it reflects how modern marketing increasingly works. Culture moves fast. Attention shifts quickly. Formats emerge, peak, and get stale.

That environment rewards teams that can notice something early, test it quickly, and decide whether it deserves more investment. That’s where AI becomes genuinely useful as a way to tighten feedback loops.

That’s what I found myself thinking about after Cannes. The AI creativity debate often focuses on the wrong end of the workflow: the output. Can AI generate ads? Can it produce content?

Those questions matter, but from where I sit, the more useful questions are operational:

• Where does AI actually help a team move faster?
• Where does it reduce friction?
• Where does it improve decision-making?
• And where does it make things worse?

That feels like a much more interesting conversation because most real adoption won’t happen through some dramatic overnight transformation. It’ll happen through smaller, quieter workflow shifts that help teams move faster without losing judgment.

That, to me, feels much closer to the future.