IBM

IBM needed Watson AI to manage 75,000+ healthcare inquiries per month with a 90% deflection target. One Primero engineered an integration that had never been built before.
90%+
AI deflection rate
125,000+
Voice calls per hour at COVID-19 peak
IBM's estimated workforce need — avoided
IBM Watson Health technology showcase at a healthcare innovation event.

The Situation

IBM's healthcare contact centre operation managed thousands of daily patient interactions spanning appointment scheduling, urgent medical coordination, and healthcare programme navigation. IBM set a clear strategic target: Watson AI should handle 90% or more of interactions without human intervention. NiCE CXone was the CCaaS environment. At the time of engagement, NiCE had no native AI capabilities.

Thus, every component of the integration would need to be built from scratch. IBM's healthcare interactions involved protected patient health information requiring HIPAA compliance. Internal data showed that 60% of patients would abandon a call if they perceived a system transfer, so the AI-to-agent handoff had to be completely invisible.

What We Did

One Primero built a custom AI orchestration layer inside NiCE that did not exist in the platform. Watson processed chat, SMS, and voice interactions. The team manually coded approximately 80% of the AI training pipeline, iterating Watson's responses across thousands of interactions to reach the 90% deflection target.

End-to-end encryption for protected health information was implemented via custom TLS SIP trunk configuration and secured REST API communication. The AI-to-agent handoff produced no audible transition. The solution deployed progressively: chat and SMS first, then voice.

What Followed

Watson handled more than 90% of interactions without human intervention. During COVID peak, the system scaled to handle 125,000+ voice calls per hour across vaccination scheduling and testing coordination across multiple US states. IBM estimated they would have needed to double their workforce without this solution.

The patterns established here — custom AI orchestration, clean handoff, HIPAA-compliant engineering — became the foundation for every subsequent AI–CCaaS integration One Primero has delivered.

Capabilities Used
AI–CCaaS Integration
HIPAA-compliant Engineering
Custom TLS SIP Configuration
AI-to-agent Handoff Design
Client
IBM
Technology / Healthcare CX · United States
Questions about a similar challenge?
Talk to the team that ran this engagement.
Talk to an Expert
AI Implementation
Ferrellgas

One Primero engineered a Cognigy–NiCE CX integration that automated self-service, payments, and agent assist — without disrupting existing workflows

Read the Full Case Study →
CCaaS Implementation & Custom Integration
Netsmart

No existing API framework. No defined endpoints. One Primero deployed CXone and co-designed the ServiceNow integration from scratch with Netsmart's IT team.

Read the Full Case Study →
CCaaS Implementation & Managed Services
Great Day Improvements

A production-blocking dialer failure resolved and a more scalable outbound operating model established across multiple brands.

Read the Full Case Study →