Recommendation screening
Summarise approved evidence for review.
EXPLAINABLE ADVISORY PILOT
A recommendation should show what it knows—and what it does not.
This live screening checks whether the minimum evidence for irrigation review is present. It never invents field measurements or prescribes water volume.
SCREENING RESULT
—confidence
Requires human review
Missing inputs
Controlled actions
Sources and protocol
NEXT STEP
Activate this workflow with your own data.
Tell the CONSAI team the geography, crop, organization and intended outcome. We will route the request to the correct pilot workflow.
EXPLAINABLE ADVISORY
Turn approved inputs into recommendations that remain open to review.
The engine screens field, weather, satellite, market and knowledge inputs, then presents visible sources, confidence limits and unresolved gaps.
Evidence boundaryPublic-safe overview: operational claims require source, owner and review status.
Source trace
Preserve why an output was produced.
Human approval
Escalate low-confidence or high-risk cases.
TRUST & GOVERNANCE
What this page proves—and what it does not.
CONSAI keeps platform activity, reviewed evidence and verified outcomes as separate states.
- Sources, freshness and review status stay visible where available.
- Platform activity is not presented as a verified outcome.
- Private records remain access-controlled.
- Human review is required before consequential action.
FREQUENTLY ASKED QUESTIONS
Questions before activation
Is this a generic chatbot?
No. It uses approved data and defined tasks.
What does confidence mean?
It describes evidence strength within a defined ceiling.
Are actions automatic?
No. Field action requires human approval.