Case study · Frontier AI lab · Computer-use models
From 200k to 1M views a post — and a pipeline of enterprise developers.
A frontier AI lab was bringing computer-use models to market. The technology was ahead of its category; the reach and the developer mindshare weren't there yet. We ran the positioning and the distribution.
~200k → ~1M views per X post.
Hackathon demand beyond capacity.
Sustained interest to program a series.
The challenge
Computer-use models only matter if developers build on them. The lab needed two things at once: a sharper public narrative, and a concrete way to put the technology in front of capable, enterprise-grade developers who could deploy it.
What we did
We owned positioning and distribution end to end:
- Rebuilt the content engine and tightened the narrative. Per-post reach on X went from roughly 200k to 1M views— a 5× lift — with no change to posting cadence.
- Organised and ran their flagship hackathon, in partnership with NVIDIA. It was oversubscribed by 300 people.
- Recruited a strong cohort of developers from large enterprises — the audience that matters most for computer-use adoption — and put the models directly in their hands.
- Turned a one-off into a program: the event generated enough sustained interest to plan 4–5 future events.
Why it worked
The technology was never the problem. Reach and the right room were. We made the lab legible to the people who could deploy it, then built the channel to reach them at scale — the same engine we run for every client.