Agri-SAGE: Simulation-Grounded Multi-Agent LLM for Context-Aware Agricultural Advisory Generation

Key findings

• ▸All three reasoning architectures (Plan-and-Solve, Reflexion, Tree of Thoughts) outperformed the static Package-of-Practice baseline across a 10-year retrospective simulation on maize in Mandya, Karnataka.
• ▸Tree of Thoughts achieved the highest 10-year average yield of 9,262 kg/ha, a gain of +1,152 kg/ha over the 8,110 kg/ha static baseline.
• ▸Plan-and-Solve and Reflexion reached similar yield ceilings (9,045 and 9,002 kg/ha respectively), but Reflexion matched that performance at substantially lower computational cost via episodic memory.
• ▸During the severe 2019 drought, Tree of Thoughts avoided crop failure by shifting the sowing date a full month earlier, outperforming the baseline by 1,384 kg/ha that year.

Abstract

Agricultural advisory systems face a fundamental tension: static agronomic guidelines offer consistent, evidence-based recommendations, yet remain blind to in-season variability and dynamic uncertainties. Recent advisory systems powered by LLMs are liable for a different risk of generating recommendations that are agronomically credible but physiologically unconvincing. Agri-SAGE is a closed-loop framework designed to resolve the above two limitations by integrating retrieval-grounded multi-agent LLM reasoning with APSIM-based biophysical simulation, to generate and validate agronomic advisories. To assess this framework, we evaluate three reasoning approaches, namely Plan-and-Solve, Tree of Thoughts, and Reflexion, over a 10-year retrospective analysis. All three significantly outperform static PoP (Package-of-Practice) baselines, with Tree of Thoughts achieving impressive peak yields. At the same time, Reflexion achieves comparable agronomic outcomes at substantially lower computational cost by leveraging cross-seasonal episodic memory.

Read the story