The marketing website, Beekeeper for fleets, the developer tools and the mobile app
Bee Maps sells map data to enterprises and dashcams to the fleets that collect it. Four surfaces, one system, one designer: a marketing website built for the 55-second scan by a technical buyer; Beekeeper, the fleet management platform that turns live devices, trips, incidents and video into something an operator can act on; the developer tools that sell the same data as an API; and the mobile app that makes every device state understandable.
A sales surface for technical buyers.
The site had one job: convince an enterprise map-data buyer in the fifty-five seconds they give a vendor. Clear promise, proof, and a path to evaluate. The narrative and the visual tone moved toward business decision-makers while keeping the brand equity earned in the contributor days.
Home, product, membership, checkout: every page built to be scanned and then read. I partnered with engineering on the purchase and account flows so a customer could buy a Bee and activate Beekeeper right after checkout, with nobody in between.
Fourteen rounds, three shown.
It started as a dark, data-first descendant of the Hivemapper site and ended as a light, product-forward page with the dashcam and the map as the heroes. The turn to light came at round eleven, when the audience research made clear that enterprise buyers read the site during the workday, next to spreadsheets. The first round, the eleventh and the last.
Buy a Bee, get a fleet.
The membership and Beekeeper pricing pages hand off to checkout, and checkout hands off to the platform. A buyer picks a region, a plan and a quantity, pays, and lands in Beekeeper with the order already tracked and setup already waiting. No sales call required to start.
An operator tool, not a dashboard for its own sake.
Beekeeper is fleet management for the companies that run Bee dashcams. It had to answer two questions fast, even when device states got messy: what is happening, and what should I do next. The work was less about adding features and more about removing ambiguity. I modelled every device and account state so the interface stays calm under complexity.
I designed all of it, desktop and mobile: checkout, onboarding, adding dashcams, tracking drivers, incidents and the AI events behind them, video retrieval, the assistant, settings, billing, orders, users and integrations. What follows is the platform, screen by screen.
01 / 08
Order confirmed
The account exists the moment the order does. Beekeeper opens with the order tracked and the rest of the product visible but locked.
01 / 07
The empty fleet
An empty state that teaches: what a driver is here, what a camera is, and the one button that starts.
01 / 06
Overview
Camera, status, last known position, and the fields the fleet chose to track.
AI events, and the video behind them.
The cameras flag events on their own: harsh braking, speeding, collisions. Beekeeper turns each one into a place on the map, a row in a list and, when the fleet needs it, the footage. Video retrieval is manual on purpose. High-resolution clips cost bandwidth and battery, so an operator asks for the minutes they need and the request is tracked like an order.
01 / 07
The incident map
Every flagged event for the fleet, clustered by place, filtered by severity and time.
Ask the fleet a question.
The assistant sits over the same data. Which drivers were speeding this week, where did the van go on Tuesday, how many incidents at the depot. Plain questions, answered with the map and the rows already on screen. The loop below is the launch piece I made for it.
01 / 08
Billing
Payment, invoices, the plan. Errors are inline and say what to fix.
The same data, sold as an API.
Beekeeper's data is also a product for developers. Keys with scopes, and a playground that runs real requests against the fleet's own data, so an engineer can see the shape of a response before writing a line. Draw a point, a circle or a polygon on the map and the request writes itself.
01 / 05
A key
Generated once, shown once, scoped to what the integration needs.
Confidence, one state at a time.
People were doing real work with the app and could not tell whether the system was behaving. I treated the interface like instrumentation: surface the right signals, explain the current state, give control without adding noise.
Two device models, several subscription scenarios, Wi-Fi or LTE, transferring or idle, searching or connected. Every combination has a clear status line and a next step. The prototype below is the app itself, cycling through the states.
Not Connected
Data
bee
Status
Tap to Connect
UPLOAD
Upload Mode
Bee LTE
Not Connected to Bee
HISTORY
Last Upload: Yesterday
–
App
Bee
Last Hour
–
57 MB
68 MB
Today
–
776 MB
826 MB
This Period
–
102 MB
451 MB
Last Period
–
2700 MB
1300 MB
Search for a place
Tap to Connect
The prototype cycles through every state. Hover to hold, pick a state, switch modes.
