Skip to content
Air Automations
All work

Radical Storage · B2C travel · 1,800+ storefronts · 2026

An autoposting agent that turned 1,800 storefronts into a live local-events calendar

Radical Storage's network of luggage-storage points lacked the team to post daily to each city's Google and Facebook pages. We built an agent to find nearby events and draft posts in each city's voice. It publishes them to Google and Facebook daily.

Radical Storage

1,800+

storefronts on a daily, event-aware posting loop

Context

Radical Storage runs Europe's largest network of luggage-storage points — partner storefronts across cities like Paris, London, Rome, Barcelona, and Milan. For a network at that scale, Google Business Profile is the foot-traffic engine: it only pays off when posts are fresh, local, and tied to what's actually happening nearby.

The problem

Doing that by hand across 1,800+ locations was unworkable. A single content marketer could touch maybe forty storefronts a week, and the posts that did go out were generic — same copy, swap the city. Conversion stayed flat and the team burned out trying to keep up.

What we built
  • Built a daily pipeline that pulls upcoming events from a curated events warehouse (PredictHQ-fed) and matches each storefront to high-attendance events within a few kilometres of its address.

  • An agent drafts event-specific copy in the Radical Storage voice and a matching image for every post, with UTM-tagged deep links so the team can attribute foot-traffic back to the work.

  • Every post goes through a single Uberall API call that publishes it to Google Business Profile, Facebook, and every other connected directory at once — no separate integration for each channel, no copy-paste.

  • A scheduler runs the whole process daily. New events trigger new posts; expired ones disappear automatically. Humans only see the dashboard when something needs a decision.

Results

1,800+

storefronts on the daily posting loop

5 cities

in the pilot — Paris, London, Rome, Barcelona, Milan

40+ h/wk

of manual posting absorbed by the agent

Stack
  • Claude
  • PredictHQ
  • Uberall API
  • FastAPI
  • Postgres
  • Railway

More on what we do for

Multi-location hospitality

See solutions