News · Google's Flow Sessions turns an artist residency into a product feedback loop

Nov, 174 min to read
AI Products

Google's Flow Sessions turns an artist residency into a product feedback loop

A two-month pilot gave a cohort of filmmakers unlimited access to Flow, then packaged what they learned into three lessons and a recruiting pitch for the next round.

What Google actually ran, not just what it wrote about

The framing is three lessons about creativity, but the operational facts are more interesting. Google kicked off Flow Sessions in September, gave a first cohort of artists unlimited access to its AI filmmaking tool Flow, and layered on mentorship and workshops over roughly two months.

That is a residency structure, not a giveaway. Unlimited access removes the usage-metering friction that normally shapes how people test a generative tool, so Google could observe what artists do when cost and quota aren't the constraint. Mentorship and workshops mean Google engineers and staff were in the room while the work happened.

The post also notes a second cohort has already started. A rolling cohort model is a deliberate choice: it keeps a steady stream of real projects flowing through the tool and produces a recurring supply of finished short films to point to.

The three lessons are also product signals

Read as customer research, the three takeaways describe where Flow's value and friction actually sit. Lesson one, 'embrace a director's mindset,' is a way of saying the tool rewards users who arrive with story, character, and cinematography decisions already made. Artist Leilanni Todd is quoted directly on this point.

The magic happens when you bring your own vision, art direction, storytelling and point of view to guide [Flow] — that's where something truly original emerges. It's less about replacing creativity and more about expanding the ways you can express it.Montana Labs

Lesson two, that 'technical know-how isn't a barrier,' is a claim about the onboarding curve. Google reports the cohort spanned all levels of technical experience and that curiosity, not skill, predicted success. For a company trying to widen Flow's addressable audience beyond VFX specialists, that is exactly the finding it needs to be true.

Lesson three, 'tell your untold stories,' points to the use cases Google wants to associate with Flow: personal, emotional, memory-driven work rather than commercial or synthetic-looking output.

Why the grandmother film and the Taiwan photos are the real argument

The two named projects do more persuasive work than any of the lessons. Chris Carboni built a film around interview recordings of his late grandmother discussing scary movies, pairing high-end visuals with her loose retelling. Katie Luo's visual poem 'The Sun Returned' transformed real photographs from a trip to see her grandparents in Taiwan into dreamlike scenes about generational love across language barriers.

Both examples anchor Flow to existing personal source material — recordings, photographs, family history — rather than to text-to-video generation from nothing. That is a specific positioning decision. It answers the common objection that generative video is soulless by showing the tool metabolizing a person's own archive into something they describe as an heirloom.

The implication: Google is using curated creative work to define what Flow is for

For teams building generative media products, the notable move here is not the technology but the go-to-market mechanism. Google is not benchmarking Flow on fidelity or speed in this post; it is defining acceptable and aspirational use through a hand-picked cohort and their finished films.

That approach shapes expectations before broad availability does. By elevating personal, story-first projects and quoting artists who frame Flow as expanding creativity rather than replacing it, Google is pre-writing the narrative that will follow the tool into wider release at flow.google.

The tradeoff is that curated residency output sets a high, mentored bar that ordinary self-serve users won't have. The honest question this post leaves open is how much of the artists' success came from the tool versus the unlimited access, workshops, and Google support that came with it.

Find this story relevant to you?

Contact us to find a unique solution

Contact us

Need an AI engineering partner that can actually build?

We help businesses integrate AI, build AI-powered products, automate high-value workflows, and modernize the software systems behind them.

Get in touch

Related reading

More analysis around product delivery, operational AI, and the systems work that makes deployment hold up in reality.

Jul, 144 min to read
AI Products

How Google DeepMind rebuilt Pelé's unfilmed 1959 goal from archives and stunt footage

Jul, 134 min to read
AI Products

Expedia's image-selection automation is the concrete piece behind its AI marketing story

Jul, 134 min to read
AI Products

ENEOS Materials built over 1,000 custom GPTs and put ChatGPT Enterprise in front of every employee