The Brief
When the lead editor at a mid-size Soho post-production house reached out, he didn't start with "we want AI." He started with a sigh.
"Every single project," he said, "we spend the first two days just organising footage. Logging clips, tagging scenes, pulling selects, generating rough transcripts. None of it is creative work. It's mechanical. And it eats into the time we should be spending on the actual edit."
His studio handles corporate films, documentaries, and branded content for agencies across London. They're a team of eight — four editors, two producers, a colourist, and a sound designer. The bottleneck wasn't talent. It was the hours of prep before talent could do its job.
What We Built
We deployed OpenClaw on a Mac Mini sitting under the lead editor's desk. Connected to their Dropbox and Frame.io accounts, the agent watches for new project folders. When raw footage lands, it gets to work.
The OpenClaw agent now handles three things that used to be done manually:
- Clip logging and tagging — analysing footage metadata, generating descriptive tags based on visual content, and sorting clips into scene folders following the studio's naming conventions.
- Transcript generation and alignment — producing time-stamped transcripts from interview footage and syncing them with the timeline for the editors to search through.
- Select pulls — flagging the strongest takes based on criteria the editors define (clarity of speech, framing, emotional tone) and compiling a "best of" shortlist for review.
The Moment It Clicked
On the third day of the deployment, the lead editor opened a new project folder expecting to spend his morning logging 14 hours of documentary rushes. Instead, he found them already organised — tagged, transcribed, and sorted into scene bins.
"I just stood there staring at my screen. It had done in four hours overnight what would have taken us a full day and a half. And it followed our naming conventions perfectly."
The agent doesn't replace the editors. It doesn't make creative decisions about pacing, music, or narrative arc. What it does is remove the mechanical drudgery that sits between receiving footage and starting the real work.
The Result
Within the first month, the studio estimated they'd reclaimed roughly two full working days per project. For a team that typically runs three to four projects simultaneously, that translates to six to eight days of editor time freed up every month.
The producers noticed something unexpected too: turnaround times dropped. Not because the editors were working faster, but because they were starting creative work sooner. The dead zone between "footage arrives" and "edit begins" shrank from days to hours.
The lead editor put it simply:
"We didn't hire AI to replace anyone. We hired it to stop wasting our best people on work that doesn't need a human brain."