Industrial Scientific has built SensAI, an AI-powered customer-support agent on Oracle Cloud Infrastructure (OCI), that has handled over 2,230 support emails automatically, saved more than 185 support hours, and driven a 30% efficiency gain while compressing response times from days to minutes for a support queue that previously processed roughly 1,200 requests per month.
"Moving from multiday email responses to automated, instant support has transformed how our technical teams operate, freeing them to focus on complex issues rather than routine inquiries," says Phanidhar Chilakapati, global data architect at Industrial Scientific.
The deployment is wired around eight OCI services working as one pipeline: a customer email lands in Salesforce and triggers an OCI Functions workflow, which coordinates with OCI Generative AI (running Cohere's R+ LLM, selected for its enterprise-grade data-privacy terms that prevent data sharing) and OCI Generative AI Agents (the RAG layer) to compose an answer grounded in OCI Object Storage, where the product-documentation knowledge base lives. Responses publish through OCI Streaming into Oracle Integration Cloud, which feeds them back into Salesforce customized to each customer's support contract. Oracle APEX renders the internal chatbot UI for support staff, while Autonomous Data Warehouse stores interaction data queryable through Select AI.
The deployment responds to a concrete operational squeeze: Industrial Scientific's support team was buried under 1,200 monthly requests on gas-detection device issues, troubleshooting, and documentation queries, with reply times often stretching to multiple days because answers required consulting multiple departments and legacy knowledge sources. Industrial Scientific is part of the Fortive corporation and a global leader in gas-detection and safety solutions for the oil and gas sector, where the Cohere selection was a deliberate data-privacy fit for customers handling safety-critical equipment documentation.
"Finding an enterprise-ready AI solution that could securely handle our documentation and improve customer service requests was game-changing," Chilakapati said. Looking ahead, Industrial Scientific is exploring chatbot-assisted support training, generative-AI-produced training materials drawn from existing documentation, and deeper Oracle APEX integration that could displace more expensive third-party chat-agent tooling. "The insights we're gaining from this system aren't just solving today's support tickets — they're helping us improve our documentation and build better training materials for tomorrow," Chilakapati added.