Rudex · 2024 – active

Face-ID and breathalyser gate for every shift

The story

The pre-shift check went from minutes to seconds, for every worker, every shift.

A purpose-built device at the gate verifies each worker's identity and sobriety in one pass, and keeps a record Rudex can stand behind.

Worth noting

· One queue at the gate instead of two
· Nobody starts a shift unless both checks pass
· The internet can go down. The gate keeps working.

The problem

Staff tested 100+ workers for alcohol before every shift. It was slow, and the test had no link to who the worker was. Identity and sobriety were two separate procedures on two separate machines, so two queues formed at the entrance, and the record of who had been tested could be disputed. In manufacturing, an intoxicated worker on the production line is more than an HR problem.

Outcome

The check before each shift takes seconds instead of minutes. Every record carries the photo from the identity check, so who was tested is no longer open to dispute. Each shift's record feeds the HR system and the company's operations system, and nothing in it can be changed after the fact.

What we built

A worker walks up to the device at the gate. It recognises the face, runs the breath test and permits or refuses entry, with no operator involved. A small computer inside the device writes every event to its own record and hands it on to the central system. On the central dashboard, HR sees attendance by shift, supervisors schedule shifts on a calendar and watch arrivals as they happen, and both can pull up a worker's or a shift's history of passed and failed checks.

What it covers

Device at the gate, face recognition, breathalyser control, offline operation and sync, central dashboard, shift calendar and attendance, photo evidence, audit trail.

How we approached it

Hardware demands a different discipline from a web app. The device has to work every shift, every day, so we designed it to work on its own and catch up with the central system when the network returns. Speed mattered as much: at 90 seconds per worker, a queue forms at the gate. Every choice, from the face recognition model to the small computer inside the device, had a reason tied to conditions on site.

Related case studies

Next case study

Twenty years of sticker albums, now an app →

Have a similar problem?

Tell us how your operation works today. We'll tell you what we'd build.

Get in touch