Skip to main content

Reasoning · AI reasoning layer

Terra Intelligence

Terra Intelligence is what makes the rest of the suite worth connecting. Sensor readings, camera detections, weather, geography, crop information and historical observations are correlated in one place, so conditions that mean little apart are read together. It produces anomalies, risk scores, predictions and recommended actions, and every one of them carries the signals it was derived from and a confidence value.

  • Signals

    • Low soil moisture
    • High temperature
    • Canopy stress in imagery

    Elevated water-stress probability

    Raised as a probability with the contributing readings attached, and routed to irrigation priorities.

  • Signals

    • High humidity duration
    • Leaf symptoms detected
    • Favourable temperature range

    Elevated disease risk

    Reported as a suspected condition requiring agronomic confirmation, never as a diagnosis.

  • Signals

    • Rising pest counts
    • Adjacent zones affected
    • Conditions favour development

    Possible infestation spread

    Direction and pace are estimated from the zones involved, and inspection is prioritised accordingly.

Capabilities

What Terra Intelligence does

The reasoning layer. Combines every signal on the farm into risks, predictions and recommendations that name their evidence.

Sensor and vision fusion
Readings and detections are resolved onto the same zones and the same timeline, which is the precondition for any correlation being meaningful.
Anomaly detection
Departures from a zone's own established behaviour are identified across every connected channel, not only the ones being watched.
Risk scoring
Water stress, disease, pest and environmental risk are each scored per zone, with the contributing signals listed against the score.
Prediction
Where current conditions have historically preceded a problem, the risk is raised in advance and expressed as a probability rather than a verdict.
Correlation across zones
Patterns appearing in neighbouring zones are linked, which is how spread is distinguished from a set of isolated events.
Explained recommendations
Every alert states what was observed, why it was raised, how confident the platform is, and what it recommends inspecting or confirming.

Data flow

What it reads, and what it hands on.

Reads from

  • Normalised sensor data
  • Computer vision detections
  • Observed and forecast weather
  • Farm geography, zones and crop records
  • Historical observations and outcomes

Produces

  • Ranked risks with confidence values
  • Anomalies and correlations
  • Predicted conditions
  • Recommended inspections and actions

Next step

Request a Farm Assessment

Every agricultural operation has a different data footprint. Terra deployments are scoped according to the farm, existing hardware and required intelligence capabilities.

The assessment establishes what your existing hardware can support before anything is committed.