Rox builds its revenue OS on a Vercel foundation
Rox is developing a revenue operating system that deploys AI agents to research prospects, generate outreach, and handle seller workflows. The company has built its application stack on Vercel from the start, using the platform's infrastructure to support weekly feature releases and global customers spanning the Middle East and Europe.
For Rox, the priority is velocity. The team needed infrastructure that would not become a bottleneck during rapid iteration cycles. Vercel's edge network and CDN provide consistent performance without additional configuration, while developer experience features such as image optimization, serverless caching, and preview environments let engineers spend less time on infrastructure and more on product development. Frontend deployments complete in 3–5 minutes, which is fast enough to push updates immediately before a customer demo. Post-build hooks also allow the team to run end-to-end tests on preview environments with minimal setup.
“Vercel has a very solid reputation in the tech ecosystem. All of the highest velocity engineering teams are using it to host their web applications.”
—Jerry Zhou, Software Engineer at Rox
Batching agent traffic with serverless functions
A central feature of Rox's platform is real-time research across thousands of prospects, which requires spinning up hundreds to thousands of AI agents simultaneously. That workload creates a technical constraint: browsers are typically limited to 100 concurrent HTTP/2 calls, and scaling requests linearly across entities leads to stalled connections.
The old approach scaled linearly with the number of entities, often causing stalled requests.
To work around this, Rox routes agent traffic through Vercel Serverless Functions, which batch calls and stream results back to the client. The result is a faster, more consistent experience for users—whether they are in Dubai or London—with research results delivered in seconds thanks to Vercel's global network and caching.
Requests are now batched through Vercel Serverless Functions.
Shipping the conversational interface with AI SDK
Rox's chat-driven interface, Rox Command, allows users to ask questions and trigger actions such as summarizing prospect interactions or drafting outreach emails. The tool is powered by the Vercel AI SDK, which handles streaming responses from large language models directly in the browser.
Preview deployments with comments enabled the team to review new features in production-like environments, shortening product feedback loops and accelerating the release of Rox Command from months to weeks.
Rox Command performs deep research and revenue-driving actions, built with the AI SDK.
Measured improvements
Since moving to Vercel, Rox has reported several concrete gains:
- Deployments in 3–5 minutes, allowing hotfixes and features to ship even during customer-facing moments
- A 2.5x reduction in P95 load speed
- Rox Command delivered in weeks rather than months, enabled by Preview Deployments and daily iteration
- Reliable multiregion performance across Europe and the Middle East
For Rox, speed is a competitive advantage rather than a convenience. The company's transformation of go-to-market workflows with AI agents depends on infrastructure that can keep pace with aggressive release schedules and global traffic.
