Two Peer-Reviewed Papers Reveal Clinical-Grade, MLR-Friendly Guardrails and Ultra-realistic Personality Design drove 35-Minute Average Conversations and 60% Weekly Retention.

WILMINGTON, Del.: #ai--Amalgam Rx, Inc., a leader in AI-powered healthcare decision-making, announced the publication of two peer-reviewed research papers validating its Medical‑Grade AI™ framework---a clinically governed system delivering safe and highly engaging digital experiences.

The studies, published in Artificial Intelligence and Applications and Human‑Centric Computing, show how Amalgam’s multi‑layered guardrails and personality‑driven AI design achieved standout results: 35-minute average sessions, 60% weekly retention, and double typical engagement.

The first study, “Taming Large Language Models for Healthcare – A Multi-layered System,” was presented at the 26th International Conference on Artificial Intelligence (ICAI 2024) and published in Artificial Intelligence and Applications. Selected from 376 submissions, this work details Amalgam’s clinical-grade guardrails for the safe, scalable deployment of Large Language Models (LLMs) in regulated healthcare environments.

This research underscores the importance of Amalgam’s Medical-Grade AI™ and aligns with a 2024 Microsoft Research study (The Illusion of Readiness). In Microsoft’s study, six leading AI models were put through a series of stress tests. They found that minor test changes dropped accuracy from 83% to 52%-showing even top-models fail nearly half the time and produce inaccurate or fabricated medical answers.

Amalgam’s second peer-reviewed paper: “Humanizing AI: Enhancing User Engagement in Health Applications with Personality-Driven AI Design,” explores how advanced LLM’s paired with human-centered design can dramatically boost engagement in digital health applications.

Amalgam’s award-winning AI chat solution engaged thousands of users over several weeks and produced the following results:

• 35-minute average sessions
• 10% exceeded one hour
• 60% weekly retention – 2x the industry baseline
• The empathetic persona drove the strongest results, averaging 41 minutes with 67% retention

While LLMs have transformed many industries, most LLMs are entrusted to self-regulate, which is insufficient for healthcare. Amalgam’s multi-layered guardrail framework sits outside the LLM. This optimizes monitoring and governs output, preventing potentially harmful information.

Additionally, such high levels of engagement have the potential to convert to outcomes in ways that include improving medication adherence and supporting patient journeys from end to end.

“If not kept in check, AI can bring chaos in healthcare-it’s like expecting obedience from an untamed beast.” said Bharath Sudharsan, chief data scientist and head of AI at Amalgam. “Our clinical-grade LLM guardrails go beyond standard safety filters, enabling our pharma partners to build safe, effective, and empathetic companion experiences.”

“At Amalgam, while we’re pushing the boundaries of innovation, our clinical and regulatory expertise keeps us grounded in patient safety,” said Ryan Sysko, chief executive officer of Amalgam Rx. “These guardrails don’t just contain AI’s risks-they unlock its potential, enabling scalable companion applications and meaningful engagement.”

About Amalgam Rx
Amalgam Rx empowers patients and providers to make the best decisions possible-within clinical workflows and in everyday life. With a modular SaMD platform, EHR solutions, and Medical-Grade AI, Amalgam partners with global life sciences companies, health plans, and providers to reimagine care delivery. Today, Amalgam’s AI-powered solutions support nearly 10 million patients across four continents and have enabled more than 70 million clinical decisions. Learn more at www.amalgamrx.com.

Fonte: Business Wire