Health Care and Social Assistance2024Generative AINLPB2B2C
Accolade

Healthcare company Accolade cuts after-call work by 50%+ with AI summarization across 100% of member interactions

Accolade implemented Observe.AI Summarization AI to automatically generate consistent, actionable after-call notes for all 14 million members, cutting after-call tasks by more than 50% and freeing Care Advocates to focus fully on member engagement.

After-Call Work Reduction50%+
Member Interactions Covered100%
4 min read

Background

Accolade's care model depends on high-quality, consistent documentation of every member interaction — both for care continuity and regulatory compliance. As the company scaled to 14 million members, the variability and administrative load of manual note-taking became a bottleneck. The goal was to automate documentation without sacrificing the accuracy and personalization that are central to Accolade's healthcare-navigation model.

What Was Implemented

  • Partnered with Observe.AI to extend existing speech analytics and transcription infrastructure with Summarization AI
  • Developed a "golden model" for note quality: defined data points, templated format, and next-steps requirements
  • Generative AI produces structured after-call summaries for 100% of member interactions
  • Summaries flow automatically into the CRM, linked to the original interaction recording for verification
  • Safeguards include the ability to listen to or view the original interaction to check any note for discrepancies

Results

After-call task time reduced by more than 50% , freeing Care Advocates for additional calls, training, or direct member engagement. 100% of member interactions are now covered with consistent, actionable AI-generated summaries. CRM records are more uniform and complete, supporting HIPAA compliance and QA. Care Advocates report reduced cognitive load and improved ability to stay present with members.

Lessons

  • Defining a "golden model" of note quality before deployment ensures generative AI produces outputs that meet real clinical and operational standards — not just grammatically coherent text
  • Linking generated summaries to the original call recording within the CRM provides a verification mechanism that builds trust in AI-generated content for compliance-sensitive environments
  • Releasing Care Advocates from administrative tasks during calls improves the quality of the human interaction itself, not just back-office efficiency

Ready to implement AI in your commerce operations?

McFadyen Digital helps teams move from case study to live implementation.

Talk to an expert →