Healthcare provider call centers get many questions about benefits, eligibility, claims, authorizations, and referrals. Many questions repeat and take a lot of time and resources. Old Interactive Voice Response (IVR) systems were made to automate these calls but often did not work well. They had trouble understanding medical terms, language differences, and spoken questions in noisy or technical places.
Humana, a big US health insurance company with over 13 million members, handled more than one million provider calls each month. About 60% of those calls asked common questions that needed clear and correct information about insurance benefits and eligibility. Their old IVR system had problems like many calls being passed to live agents, longer wait times, and high costs. Many providers wanted to talk to a real person because the automated system did not work well. This showed the need for a better AI system.
To fix these problems, Humana worked with IBM Watson to make a conversational AI Voice Agent for healthcare provider calls. This AI uses special training and model improvement:
The AI system started in April 2019 and keeps improving as more providers use it. It answers over 7,000 voice calls daily from 120 different providers, handling many routine questions fast and well.
This AI agent has cut Humana’s call handling costs to about one-third of what the old IVR system cost. The AI also nearly doubled the number of calls answered automatically. Calls now take about two minutes without waiting for a real person to help.
In real life, this means healthcare office managers and IT staff in the US have fewer routine calls to handle. Human agents can focus on harder questions that need personal attention. This makes providers happier because they get quick answers and fewer call transfers or delays.
The AI also gives detailed replies like confirming exact co-pays for services such as chiropractic care. Tasks that used to take a long time or require faxes now happen instantly through the automated voice system.
Besides saving money and time, Humana’s AI collects data about provider calls. This data helps make the system better. Sara Hines, Humana’s Director of Provider Experience and Connectivity, said the AI has helped improve how routine provider calls are handled.
Building on this success, healthcare IT managers and office leaders can use AI for more than just answering calls. AI systems like Simbo AI help with phone automation and answering services for patient scheduling, appointment reminders, and insurance checks.
AI automation gives these advantages:
Together, AI phone automation and workflow tools can improve productivity and the quality of communication between patients and providers across healthcare practices nationwide.
The US healthcare system faces special challenges like complex insurance rules, many types of providers, and strict safety standards. These need accurate and secure communication tools built for US medical settings.
Humana’s work with IBM Watson shows how American healthcare can build AI models for:
Humana’s AI project took three years and kept improving based on real experience. This way, the AI learned to handle real provider needs and tricky situations like rare terms or hard conversation parts.
The project led to several pending patents and helped create the AI model used in IBM’s watsonx Assistant for Voice Interaction. This points to new standard AI tools that other healthcare groups can use, making advanced voice technology more available.
Providers have mostly reacted positively to the AI. They get fast access to key coverage information and face fewer system errors or call transfers. This shows growing trust in AI tools at healthcare offices. It also moves workflows toward more tech help that can reach beyond insurance to clinical and admin tasks.
Many providers in rural or underserved parts of the US have bad internet. This makes using complex AI hard. AI model improvement for low-bandwidth is needed here.
IBM Watson made voice models that use less data and work even when the voice input is weak or partly lost. This helps small clinics or places with limited IT still have good AI help for providers.
Making AI work well in all areas helps healthcare automation reach more places, from big city hospitals to small rural clinics.
Humana’s AI voice assistant shows that custom AI training and model fixing can improve understanding of healthcare terms and call center work, especially in tough places like low-bandwidth areas. This gives US healthcare managers and owners useful tools to cut costs, make providers happier, and run routine calls better.
Using AI phone automation with workflow improvements helps healthcare offices deal with common questions on insurance, benefits, and authorizations more smoothly. Providers get quick and correct answers with fewer call moves, making work flow better and staff time used well.
As AI tech keeps growing with ongoing training and real case feedback, healthcare groups should see better communication, smoother operations, and lower costs. This fits well with goals to improve healthcare work in the United States.
Humana’s legacy IVR system transferred too many calls to human agents, incurring high costs and lowering customer satisfaction. Over 60% of calls were routine pre-service questions with well-defined answers, yet providers preferred speaking to live agents, increasing operational expenses.
The PSI team aimed to reduce costly pre-service calls and improve provider experience by implementing a solution that could efficiently handle routine inquiries about health plan benefits, eligibility, authorizations, and referrals without human agent intervention.
Humana partnered with IBM Watson, specifically IBM Watson® Assistant for Voice, combined with IBM Cloud® infrastructure, to build an AI-powered conversational voice agent capable of handling provider inquiries autonomously.
The Voice Agent intelligently understands caller intent, verifies system access permissions, and provides specific, context-sensitive insurance information such as eligibility, benefits, claims, authorizations, and referrals quickly and accurately, minimizing the need for live agent interaction.
Through speech customization with seven language models and two acoustic models, the system achieves an average 90%–95% sentence accuracy in understanding provider input, enhancing response precision in complex healthcare queries.
The Watson-based solution handles calls at about one-third the cost of the previous system and nearly doubles the response rate compared to the older automated IVR, demonstrating both economic and operational efficiency improvements.
The IBM Watson Expert Services Lab enhanced Watson’s healthcare terminology comprehension and optimized performance for low-bandwidth call centers through custom training and model refinement, contributing to patented innovations and published methodologies.
The Voice Agent processes over 7,000 daily voice calls from around 120 providers, with overwhelmingly positive user feedback indicating improved self-service capabilities and faster access to insurance information without waiting for representatives.
The assistant can address specific sub-intents such as verifying exact co-pays for services like chiropractic visits, which previously would have required extensive manual support or generated lengthy fax responses under the old IVR system.
Humana views AI as an evolving tool that, beyond current improvements in pre-service inquiries, has vast potential to further enhance provider communication, operational efficiencies, and customer care, reflecting an ongoing commitment to innovation driven by AI advancements.
Simbo is using AI agents to automate all phone related workflows. Connect now to know more.
Schedule AI Agents Demo Now© Since 2019, Simbo AI.
