IGAD — AI-powered agricultural intelligence & proposal generation

The problem

Securing funding for the IGAD region means writing strong, well-evidenced proposals against each call — fast. Teams did it by hand, reusing past proposals and prior experience and re-shaping them for every new opportunity. It was slow, and quality depended on who was writing.

The approach

We built an AI-Powered Agricultural Intelligence Hub for the region's experts and policy makers. Its flagship tool, the Proposal Writer, takes prior proposals, documented experience and the target call for proposals, then uses several models on AWS Bedrock to analyze the inputs and produce a documented, detailed draft — with section-by-section AI feedback (scored excellent / good / needs improvement, with concrete suggestions). A growing suite around it adds a Newsletter Generator, Report Generator and Policy Analyzer.

Architecture

• AWS Bedrock hosts the models; a multi-step pipeline analyzes prior proposals, documented experience and the target call for proposals.
• RAG grounds the drafting in the uploaded sources before the section-by-section scoring pass.
• A Node.js + Angular + MySQL application on AWS serves the tool suite.

The outcome

• Documented proposal drafts produced in a fraction of the manual time.
• A consistent quality bar via automated, section-level review.
• A reusable AI toolset for the IGAD region rather than a one-off script.

Public documentation: the IGAD tool docs.