Mister Spex automates 70% of caller verifications and 52% of order-status inquiries with AI Phone Agent
Europe's leading online optician deployed a Cognigy AI Phone Agent to handle identity verification and WISMO calls, automating 70% of verifications and 52% of order-status inquiries within three months of go-live.
Background
Mister Spex built a strong forecasting model for call volume but persistently struggled with unpredictable spikes due to sick leave and unforeseen demand. Extended hold times conflicted with its customer-centric service strategy. The decision was made to automate the phone channel — the most challenging communication channel — rather than introduce an entirely new channel, because data showed a high proportion of incoming calls were simple and repetitive.
What Was Implemented
- Cognigy AI Phone Agent deployed as the first point of contact for all incoming calls
- Numeric identity verification (house number + postal code) replacing name/email verification
- CRM integration (Microsoft) enabling real-time order status retrieval and automated follow-up email dispatch with tracking numbers
- Genesys contact center integration for intelligent routing to specialist agent teams
- Full call transcription and reporting for continuous improvement
Results
Within three months of go-live: 70% caller verification rate (fully handled by AI), 52% of WISMO inquiries fully automated , 88% of return label requests automated , 96% intelligent routing accuracy (up from 68% with legacy IVR), and 30 seconds saved per call across all interactions. All figures are vendor-reported via the Cognigy case study.
Lessons
- Automating an existing high-volume channel (phone) with AI yields faster ROI than introducing new channels, because call drivers are already understood.
- Numeric authentication outperforms name/email verification over voice, reducing errors and improving automation confidence.
- Reporting full call transcripts enables rapid identification and correction of routing errors, accelerating continuous improvement post-launch.