Hospitality
Restaurant group automates daily CCTV health checks across every shop
A quick-service restaurant group's IT team checked every shop's CCTV system by hand — logging into each shop's recorder, checking the hard disks and remaining recording time, then eyeballing every camera's image for faults. We built a robot that tours the whole shop list daily: it logs in, verifies disk health and recording duration, snapshots every camera, flags all-black images, and scores each snapshot against its reference photo — then files a per-shop, per-camera summary and emails the result.
What we used: RPA Development, Process Automation
1The problem
- IT staff logged into each shop's CCTV recorder by hand to check hard-disk status and remaining recording time, then eyeballed every camera's image — tedious, low-value work.
- Done manually, the checks ran far less often than the team wanted; a dead camera or degrading disk could go unnoticed for days.
- Two generations of recorder hardware were in service, each with a different interface and different pass criteria.
- One full-time staff member was tied up daily, and results lived in scattered notes rather than one report.
2What we did
- 1An unattended robot works through the consolidated shop list every day — shop name, code, camera count, and recorder generation drive every step.
- 2It logs into the CCTV management client shop by shop, retrying login errors and marking unreachable shops for follow-up instead of stalling the run.
- 3Disk health is verified per recorder generation — enabled-status across all disks on the new hardware, SMART status on the old — and graded OK or Failed.
- 4Remaining recording time is read from each recorder with OCR, then graded Normal, Abnormal, or Exception against the minimum recording time.
- 5Every camera is snapshotted: fully black frames are counted, and each image is scored against its reference photo with a distortion value.
- 6A summary workbook records disk status and recording time per shop, plus black-frame counts and distortion values per camera.
- 7A completion email closes every run; anything unexpected escalates by email — no shop fails silently.
3Inside the Platform
Get results like this
We’ll assess your process and give you a straight answer on what automation could return.