Creative GenAI
Key Takeaway: Everyone was building AI tools. We had to build AI tools and rewire the team using them at the same time.
Creative GenAI Platform Liftoff, 2025 to present.
By 2023, it was clear that generative AI had the potential to replace entire features and workflows. The question was whether you built that replacement, or waited for someone else to.
At Liftoff, we didn't wait and we also didn't pretend the transition was simple. Building a generative AI creative platform meant solving three problems simultaneously: the technology, the customer, and the team. Most conversations about GenAI in advertising focus on the first. The second and third are where the effort was more critical. And of those two, the harder one was always the team.
The technology can be iterated. A person who has decided that AI is a threat to their identity and their craft is a much harder problem to solve. Getting that wrong early sets the tone for everything that follows.
Two customer paths
From the outset, the platform had to serve two fundamentally different kinds of customer, and serve them well.
The first path was premium. Large advertisers with high creative expectations where quality and brand fidelity were non-negotiable. For these customers, the work remained largely hand-crafted, with AI introduced selectively as a co-pilot in specific areas: ideation, rapid iteration, localisation, asset adaptation. The human creative remained in the loop at every meaningful decision point. Introducing AI here was less about replacing production and more about extending what the team could do without increasing headcount.
The second path was full AI. A separate tier of advertiser where the creative brief goes in and campaign-ready assets come out, with the workflow managed internally by Liftoff rather than handed to the advertiser to operate themselves. This is not self-serve in the conventional sense, as the tooling is not yet in the hands of the customer. The intelligence, the quality control, and the operational responsibility sit with the team, with output being AI-native in a managed service while we build the end-to-end.
These two paths required different things at almost every level: different quality thresholds, different feedback loops, different definitions of done, and different conversations with customers about what AI could and could not yet do reliably. Holding both simultaneously, without allowing the constraints of one to compromise the ambitions of the other, was one of the defining challenges of building the platform.
Changing the team
The technology problem and the people problem arrived at the same time, and could not be solved in sequence.
The creative team at Liftoff had been built for a manual production model. Skilled designers, animators, and creative strategists who knew how to make high-performing mobile ads by hand. With the integration of Generative AI, it didn't make expertise irrelevant but it changed how it was applied.
The work of this period was helping a team of creative specialists become AI-native without losing what made them excellent. That meant being honest about what was changing, building new workflows that kept human creative judgment at the centre, and redefining what quality meant when output was generated rather than crafted.
The goal was to make the team genuinely good at using Gen AI. Some people got there quickly and others took longer. Everyone is still adapting, and that is the honest version of what a transition like this looks like from the inside.
What the platform does
The platform spans the full creative lifecycle, from brief to live asset, across three primary format areas: HTML end card generation, AI-powered video production, and banner automation for gaming and brand advertisers.
The commercial framework I own sits across both customer paths: defining how the platform is packaged and priced for full-service clients, how self-serve access is structured and supported, and how creative cycle time, the north star metric for the platform, is improved across both.
The mandate is to make AI-native ad production a genuine, repeatable capability for Liftoff's customers. Not a pilot programme or a proof of concept, but something that works, at scale, every time.
Where it sits in the longer arc
This role is, in some ways, the culmination of everything that came before it. The systems thinking from Dynamic Templates and Adaptive Creative and the operational muscle and curiosity from Vungle Creative Labs. The cross-functional leadership from Chief of Staff and the format expertise from UX and Formats Design.
But the most relevant experience has been the fifteen years spent understanding how creative people think, what threatens them, and what makes them do their best work.