Multi-channel AI Agent for personalized appointments in Healthcare

What does the company do?

The US-based healthcare company has a mission to provide high-quality, affordable, and easy-to-understand healthcare plans for seniors. It specializes in Medicare Advantage offerings and leverages advanced technology to enhance healthcare delivery. Operating across multiple states in the United States, this organization serves over 100,000 members, reflecting its expanding market presence. By the end of that year, the organization employed over 550 individuals, and it maintains a public listing on the NASDAQ.

How does Vstorm cooperate with the client?

Challenge

As the organization expanded, the need for scalable, personalized solutions became increasingly urgent to improve efficiency across multiple clinics and enhance communication with senior patients. They sought a consulting and engineering partner capable of delivering AI Agents and integrating them into existing workflows and healthcare systems.

The overarching goal was to create an AI Agent that would, in a personalized manner, analyze a patient’s entire record, including medical history, physician notes, active health issues, visit history, personal details, and current medications – while also gathering any additional information directly from the patient using multi-channel methods tailored to older adults. At the same time, the solution would continue building a robust data foundation for future use, all in compliance with HIPAA standards.

How did we approach?

• Incremental rollout. To avoid technical debt, we began with a Proof of Concept (PoC) tested by a select group of doctors. They were impressed by how much time it saved, validating our approach before scaling up.
• Semi-manual approach. In line with our methodology, we created the initial version of the AI Agent using a semi-manual approach. This ensured doctors maintained full oversight and remained actively involved throughout the process. By doing so, both Vstorm and the medical staff could identify new exceptions as they arose, continuously refine the AI Agent, and gradually transition it toward greater automation, tackling challenges related to workflows, AI capabilities, integrations, and change management along the way.
• Tailored AI integration. Rather than relying on generic or off-the-shelf platforms, Vstorm focused on creating a custom AI solution. This allowed the AI Agent to integrate seamlessly with various internal systems, such as clinic management and physician management platforms, while aligning with unique workflows and processes.
• Seamless system integration. To minimize manual data entry and ensure efficient data flow, we maintained compatibility with the healthcare organization’s core platforms, eliminating many of the typical integration hurdles.
• Self-learning system. After each appointment, the patient’s records are automatically updated and integrated into a comprehensive health history, ensuring all relevant data remains in one centralized place.
• RAG mechanism. By leveraging retrieval-augmented generation, the AI Agent can ask patient-specific, personalized questions based on diverse data categories and the patient’s complete medical history without hallucinations.
• Multi-channel communication. We implemented SMS and voice call features designed specifically for seniors who may be less familiar with technology.

I love the summaries. I think it is amazing, and I am very impressed with this tool.

End user (doctor) of an AI Agent

Results

We developed an AI Agent dedicated to the pre-appointment phase, enabling patients to share critical updates and concerns well before seeing a doctor. By using voice communication, the process feels more natural and accessible for older adults compared to traditional forms or questionnaires. Ensuring the solution met strict healthcare compliance requirements while delighting patients and care teams alike.

All of these efforts saved each doctor over five hours per week by automating information collection and building a robust data foundation for further personalization, rather than relying on generalized approaches. In addition, patient engagement rose by more than 20% thanks to the personalized, accessible communication options available whenever they needed them.