MetadevXInnovate. Code. Create.MetadevXInnovate. Code. Create.

(01) The challenge

Sensors in the ground, decisions still by instinct

TaniSmart had deployed soil and weather sensors across commercial farms, but the data ended up in a portal that farm supervisors opened once a week. Irrigation and fertiliser decisions were still made by walking the rows.

  • · Sensor coverage on 100% of pilot farms, usage under 20%
  • · No alerting when moisture crossed a threshold
  • · Supervisors on low-end Android devices in direct sunlight
TaniSmart — The challenge

(02) Approach

Built for the edge of the field

We redesigned around the supervisor's day: a high-contrast mobile view that reads in sunlight, a single alert stream in simple Arabic and English, and offline tolerance for the parts of a farm with no signal.

  • · Threshold alerts delivered as one action, not a dashboard
  • · Bilingual, high-contrast UI tested on-site in daylight
  • · Store-and-forward sync for patchy rural connectivity

(03) Outcome

Water use down 24% on pilot farms

Once alerts reached the person holding the irrigation valve, the sensor data started paying for itself. Pilot farms cut water use by a quarter and reported fewer crop-stress incidents through the summer season.

-24%

Irrigation water use

78%

Daily active supervisors

2

Languages at parity

(04) What we did

We went to the farms before we wrote the spec

Two weeks on site changed the product: the portal idea died, the alert stream replaced it, and the device matrix we targeted came from what supervisors actually carried in their pockets.

  • · On-site research across five commercial farms
  • · Alert model designed with agronomists
  • · Pipeline: sensor ingestion, thresholds, bilingual delivery
  • · Field trials with supervisors shaping each release
Next projectRyzen Drone→