ColdTrace AI began as a question-and-answer chatbot. The opportunity was to make it more operational: helping health teams understand equipment issues, find relevant guidance, report faults, request support and move into the right ColdTrace workflow.

I led the product and interaction design, working with two Product Managers and an AI Engineer. Together, we mapped workflows across equipment monitoring, maintenance, training, service requests and technician support, then translated them into a reusable assistant model: intent, context, guidance, action and escalation.

The resulting beta and prototype work connected contextual AI support across ColdTrace and WhatsApp. It introduced guided prompts, language and location handling, issue reporting, service escalation and technician handoff, while giving each channel a clear role based on the depth and complexity of the task.

‍