Smart farming
for Thai rice growers
with AI

From a real field to a redesign

From a real field to a redesign — Gen 3 → Gen 4

This isn't a lab product — every Gen 4 decision traces to something the field taught us.

Gen 3 · Songkhla 2026

What the data told us

9.1%Gen 3 average uptime
across the first wet season

• Electronics flooded as water levels rose in the paddy
• Battery ran flat during several days of continuous rain
• Intermittent data gaps broke the continuity of analysis

Gen 4 · Redesigned

Redesigned from the evidence

Gen 3 problem		Gen 4 fix
Electronics submerged in floodwater	→	Control box raised to 1.2 m, IP68-sealed
Battery drained during prolonged rains	→	Bifacial solar + 5–11 days of backup
Uptime of just 9.1%	→	Redesigned power-backup architecture

1. 2026 · Q1
   Gen 3 deployed at Ranot, Songkhla

   A 36-rai prototype station collecting real data through the first wet season

2. Problems found
   Flooding · submerged electronics · dead battery

   Uptime dropped to 9.1% — data lost across many periods

3. The fix
   Hexagonal prism mast + raised foundation

   Electronics moved up high, IP68-sealed throughout

4. 2026 · Q4
   Gen 4 certified by a professional engineer (PE-signed)

   Structure and electrical systems certified by a licensed engineer

Only the pesticide that's actually needed

Only the pesticide that's actually needed

With an image library of over 7,000 pests, RiceGuard identifies the right species and recommends the right dose — right pest, right dose, right time.

Ranot, Songkhla — what the soil revealed

Ranot, Songkhla — what the soil revealed

Sensors found soil as acidic as pH 4.7 — invisible to the naked eye. Targeted, spot-specific liming cut lime use by 57%.

pH 4.7

Most acidic spot

57%

Dolomite lime saved

36 Rai

Trial area