A U.S. agricultural cooperative managing multi-site chemical mixing operations came to Armakuni for aI-powered proactive chemical safety monitoring and natural-language chatbot interface for SDS data querying.
The challenge
Scale: 142 Safety Data Sheet (SDS) files across chemical inventory; multi-user workforce handling chemical mixing daily.
What we built
Armakuni delivered aI-powered proactive chemical safety monitoring and natural-language chatbot interface for SDS data querying.
The outcomes
5 measurable outcomes shipped across the engagement. The ones that moved the business the most:
142 SDS documents structured and made queryable. Every chemical safety document the cooperative relies on is now a unified, AI-queryable knowledge base, so workers retrieve hazard and compatibility information in seconds rather than minutes.
Three-tier safety decision output validated end to end. The monitoring workflow produces reliable green, yellow, and red output from screenshot inputs, with yellow flagging specific hazards and red surfacing corrective actions, accuracy depending on the SDS data quality the client provides.
Natural-language chemical safety queries resolved from verified sources. The chatbot retrieves and synthesizes SDS data to answer domain-specific questions accurately, giving operators actionable guidance without locating or reading source documents manually.
Production-ready AWS architecture validated at the PoC stage. The deployment supports a direct path to production without a rebuild, with CI/CD pipelines, multi-AZ availability, security tooling, and auto-scaling in place from the first delivery.
Foundation established for a multi-phase production system. The PoC is the entry point for a broader buildout, with live real-time monitoring, full integration with existing mixing software, and expanded AI capabilities documented as next steps.
Built on AWS
The production environment runs on 16 first-party AWS services. Delivered under our AWS AI Services, Cloud Native Development competencies.
