Medical Device Manufacturer resolves 85% of AI-handled order emails with no follow-up needed
fully resolved by the AI agent
SLAs for response times
CSAT scores from support inquiries
Every support email was handled manually: look up an order, check its status, and write a reply. Leaving lettle time for complex issues or selling.
A Quiq AI agent that reads an order inquiry, look up the actual order status, and either answer it outright or draft a response for a human agent to send.
85%+ full resolution rate on emails the AI agent replies to directly, with no follow-up email required from the customer
The Challenge
- Manual, one-by-one handling of order and shipment inquiries. Every question came into a support inbox and had to be read, researched, and answered by a person, with no way to triage the easy ones from the hard ones automatically.
- Missed SLAs. The volume of these routine emails outpaced what the support team could handle by hand, and response times slipped.
- Declining CSAT. As response times suffered under the manual workload, customer satisfaction scores declined.
- Sales team members pulled into support work. When the support queue backed up, sales staff were pulled in to help answer inquiries, time that came directly out of selling.
How Quiq was deployed
Quiq’s Email AI agent handles their order volume through three tiers of automation, matched to how confident the AI agent is in its answer:
- Full resolution: The AI agent reads the inquiry, looks up the order in their Oracle environment, and replies directly with the order status. When it resolves the request, it closes the Oracle incident automatically, so nothing sits open waiting on a person.
- Drafted response: When the AI agent doesn’t have enough information to fully resolve an inquiry, it drafts a response and routes it to a human agent to review and send, rather than starting the agent from a blank page.
- Human handoff: Requests outside the AI agent’s confidence threshold go to a person with no automated response attempted.
- Oracle integration: The AI agent works directly inside their existing Oracle environment rather than as a separate system the team has to check.
- Parallel-run validation before go-live: Before the AI agent replied to a single customer, it ran silently alongside human agents. Quiq reviewed its draft responses and refined them, so the system was only turned on for real customers once its answers held up under review.
What changed after launch
For the support and sales teams, the AI agent changed what a routine order inquiry means. Instead of every question landing in a queue for a person to research and answer from scratch, most of them are now either resolved outright or arrive to a human agent as a drafted response ready for review, cutting out the step where an agent starts from nothing.
That shift also freed sales team members who had been pulled into support work to get back to selling, rather than splitting their time between the two.
Results/ROI
By replacing fully manual handling of order and shipment inquiries with a tiered AI agent grounded in its own confidence, they turned one of its highest-volume, lowest-value email categories into something that mostly resolves itself.
- 85%+ full resolution rate on emails the AI agent replies to directly, with no follow-up email required from the customer
- Sales team time recovered from support tasks back to selling