AI video director Billy Boman on making Super Bowl ads, directing AI commercials for Taylor Swift and Lewis Capaldi, and having his own Hollywood-scale sign

Summary

Billy Boman is a Stockholm-based AI creative director running a lean production studio with teams of four to seven artists. He has directed branded content for Taylor Swift, Lewis Capaldi (with Google, YouTube, and Universal Music Group), and a Super Bowl commercial for a major airline.

Boman pivoted into AI video around 2022 when early models could generate convincing hovering sneakers. By 2023, spec ads he posted on LinkedIn attracted inbound from AI agencies including Monks. He was working remotely from Sweden for a German AI startup at the time; when they pivoted, he went independent. Agency relationships became the channel for all subsequent work.

Production tools

MidJourney dominated image work through 2024. On video, Google's Veo suite became industry standard before Kling 3 and Sea Dance 2 (from Dreamina) took over as the current benchmark. Boman had early access to Sea Dance through Dreamina's creative partnership program, avoiding queue and rate limits that have frustrated other practitioners. He notes Sea Dance's quality likely traces to its training data, drawn from Hollywood film libraries rather than YouTube clips, though he flags that as inference.

Against cost-cutting

Boman's core argument is that brands cutting budgets in response to AI will lose to competitors who keep budgets fixed and use AI to multiply creative output. More variations, more channels, more ambitious concepts become possible. One brief asked for a girl glamping in Argentina feeding a protein bar to a twerking capybara. Flying a crew to South America was never realistic, nor was wrestling an unpredictable animal on set.

For the Super Bowl airline commercial, the approach was hybrid: human talent shot on green screen with AI-generated environments for backgrounds and relighting. This combination of AI and live action, not one replacing the other, is the model Boman sees going forward.

Project timelines have compressed, but production still takes a month to six weeks from brief to delivery on a proper 30-second spot. The real cost is human hours, not compute credits. Casting decisions—whether to use AI-generated faces or real talent—require early commitment and determine everything downstream.

Building a team

Finding people remains the biggest constraint. Boman teaches at Berghs School of Communication in Stockholm, which runs what may be the world's only government-funded one-year classroom AI program, and uses it as a recruiting funnel. The studio is him, a co-founder, and two assistants. He is building slowly, noting that AI-native students often need supplemental training in art direction, client management, and how to take a brief.

The Fiverr partnership, where Boman is featured on a curated hub of vetted AI commercial directors, addresses a market gap. Agencies trying to hire AI talent through Instagram DMs frequently get burned. A small, curated list of production-ready practitioners solves discovery for both sides.

Taste and differentiation

The number of genuine creative tastemakers stays small regardless of the toolset, Boman agrees. Perhaps 100 globally, possibly fewer. The floor has risen sharply: anyone can now generate a convincing fight scene. But the ceiling has also risen, making genuine differentiation harder. Craft, taste, and the ability to art-direct at a cinematic level separate commercial-grade work from competent amateur output. An anthropology student with unusual vocabulary and distinct visual instincts can get results a trained filmmaker misses entirely.

On AI writing scripts and originating ideas, Boman is direct: that output is not yet good enough, and those tasks remain the part of the job he wants to keep.