By Zaffar Khan

As use cases for and the acceptance of GenAI continue to grow, health plans have a unique opportunity to not only push the envelope but rip it wide open. 

The next step in the evolution of conversational AI is GenAI voice agents. These agents engage in fluid and natural conversations and better understand nuances in human language. This leads to more accurate and contextually appropriate responses. Importantly, they can generate more human-like speech with variations in tone, intonation, and emphasis, making interactions more engaging. 

Building a GenAI Voice Agent

Launching a GenAI voice agent, or Agentic AI, offers health plans a multitude of possibilities to design and implement the technology to showcase the organization’s dedication to customer service, continuous improvement, and much more.

These are a few ideas that demonstrate the different ways to train the technology and the resulting organizational advancements:

Custom healthcare data: Voice AI agents powered by large language models (LLM) and small language models (SLM) provide the ability to fine-tune general-purpose models. In healthcare, custom vocabulary and medical terminologies can improve model performance and accuracy. The model can learn from ICD-10, SNOMED CT, and RX Norm, to name a few.

Empathetic responses: LLM powered AI agents can generate responses that acknowledge and validate the member’s feelings. This includes phrases like “I understand this must be difficult for you,” or “It sounds like you’re going through a lot right now.”  

Tone and voice modulation: The AI’s voice changes to convey warmth and concern, including adjusting the pace of speech, intonation, and overall tone. This creates a more comforting and human-like interaction.  

Privacy and data security: Open source LLMs and SLMs provide the flexibility to be hosted on-premise or in the cloud with required information security while handling sensitive patient data with strict adherence to privacy regulations like HIPAA. 

Each process applied to GenAI voice agents—to varying degrees—has led to something of a revolution in the healthcare contact center field “Generative AI has advanced significantly in recent years. It began with rudimentary algorithms that could only generate limited and predictable patterns and has progressed to more complicated tasks such as replicating human-like conversational abilities and making realistic visuals and sounds,” according to the International Research Journal on Advanced Engineering Hub. 

There is a movement underway thanks to this evolution to supply chatbots with “human” voices. It turns out, most people can’t tell the difference. 

A survey by Podcastle found that people “don’t know the difference between AI and humans – 2 in 3 incorrectly guessed that a human voice was AI.” In addition, 53% of those surveyed said they have a positive or neutral view of “AI voice technology.”

While people may be willing to converse with GenAI to get details on repairing their car, for example, having a similar interaction as a health plan member will be quite different. But the outcome: imparting timely, accurate information should be the same.

Improving Member Engagement

A GenAI voice agent can support member engagement, leading to improved health outcomes. Its job of communicating with members can help health plans accomplish several goals related to improving operations through personalized health coaching and guidance:

• Tailored interventions a real-time support: AI Voice agents can deliver personalized health coaching based on individual member needs, risk factors, and preferences. Members can receive immediate support and guidance, such as reminders to take medications, tips for healthy lifestyle changes, or advice on managing chronic conditions

• Health risk assessment (HRA): Engage members with care coordination and enroll in CMS-driven risk assessment programs  

• Survey and feedback collection on QAA (Quality of care, Attitude of provider, and Access to care) issues: Automated Voice AI surveys can gather valuable member feedback on service quality, physician interactions, and overall satisfaction with health plans. 

• Health check-ins and educational outreach: AI agents can conduct routine health check-ins, inquiring about symptoms, medication adherence, overall well-being, and personalized health education messages, explaining complex medical information in easy-to-understand terms.

Simply put, the GenAI agent calls the member to ask if they have time to complete an HRA and collect the information through a series of questions determined by the health plan. (Members called are based on CMS-guided HRA questions and frameworks—diagnosis of a chronic condition or frequent emergency department use, for example—as stipulated by the health plan.) also ensuring personalized and empathetic interactions. 

Voice-enabled GenAI agents can help health plans maintain relationships with members while gathering essential information that can improve care coordination, care delivery, and overall health and wellness. 

GenAI voice agents are poised to revolutionize outbound calling in the healthcare industry. By leveraging a variety of advanced LLMs, AI-powered voice agents can significantly improve efficiency, member engagement, overall operational outcomes, and understanding of healthcare costs. 

Zaffar Khan is Associate Vice President of Business Transformation & Generative AI Solutions at Sagility.

This article originally appeared in Healthcare Contact Center Times.