Applied Expertise
Challenge
A leading healthcare organization sought to improve its New Patient Experience but was held back by an outdated and inefficient Quality Assurance (QA) process. Supervisors manually reviewed fewer than 10% of calls, spending hours documenting feedback that wouldn’t reach service agents for over three months. This delayed response impeded efforts to improve the quality of service agents’ interactions with patients who were first-time callers. Without timely, actionable insights, the team struggled to improve agent performance, deliver consistent experiences, or provide meaningful cross-training opportunities.
Solution
SEI partnered with the client to design and implement an AI-powered solution that automated QA processes and enabled real-time insights from patient calls. By integrating generative AI and Natural Language Processing (NLP) capabilities, we:
- Converted 100% of voice calls into transcripts, enhanced with metadata and tags
- Used NLP and generative AI to summarize key details by individual agent and group, making insights easily accessible
- Developed a CoPilot interface to surface real-time feedback, monitor call quality, automate scoring, and uncover cross-training opportunities
- Automated both quantitative QA checks (e.g., completeness, key criteria scoring, and conversation flow) and qualitative insights (e.g., empathy detection, call reasons, sentiment, and resolution outcomes)
- Implemented sentiment analysis to monitor emotional shifts during calls, with a focus on transitioning patients from fear to hope
Results
The impact was immediate and far-reaching:
- Automated 100% of QA reviews, cutting feedback cycles from 3 months to under 24 hours
- Equipped supervisors to deliver real-time feedback and coaching, significantly improving call quality and empathy
- Delivered rich analytics across service agent interactions, enhancing decision-making and performance monitoring
- Deployed real-time sentiment tracking, improving service outcomes and patient engagement
- Led the end-to-end transformation, including proof-of-concept, model training, prompt engineering, deployment, and change management
