Inside Naver's 1784, a 36-level office tower in Seoul that has ~4,500 employees working with 120+ robots and serves as a testbed for various AI technologies
Saritha Rai / Bloomberg : X: @saritharai X: Saritha Rai / @saritharai : In Seoul, I found a glimpse of our future with robots At tech giant Naver's HQ, 4,500 pple work with 120 robots. Robots glide around bringing coffee, lunch, purchases S.Korea factories have record 1,012 robots to 10,000 workers @business @technology https://www.bloomberg.com/... #AI
Context & Ripple Effects
Naver’s workplace robotics effort has moved from its earlier experiments with roughly 100 office robots to a live operating environment that combines employee services with AI testing. The company had also signaled a path beyond internal use by planning to offer its Rookie helper robots to other companies after extended testing of the office-helper program.
That makes 1784 more than a showcase: it is a controlled site where Naver can observe how robots fit into routine office workflows at meaningful employee scale before wider deployment.
First-order effects
- Employees at 1784 receive on-site delivery services from a robot fleet, while Naver gains continuous real-world feedback on robot operations and AI-assisted workplace tasks.
- Naver can validate its office-robot product in its own building, reducing the gap between prototype testing and the external rollout previously contemplated for Rookie.
Second-order effects
- A proven internal deployment gives Naver a stronger reference case for selling workplace robots to companies that need evidence of reliability and employee acceptance, not just a demonstration.
- Office-service automation shifts competition toward integrated systems: robots must work with building operations, deliveries, and the software layer that assigns and monitors tasks.
Third-order effects
- If deployments like this scale beyond a single headquarters, office automation will increasingly be judged as a workflow product rather than a standalone robot purchase—linking physical machines to the same operational software used by workers.
- The model also suggests that large employers with their own facilities can become testbeds and early customers for embodied AI, though broader adoption will depend on whether the operating benefits transfer to less-controlled workplaces.
The trend: Naver’s 1784 is part of the shift from experimental office robots toward workflow-native, AI-managed physical services in everyday workplaces.