Hear what Get Well’s Senior Vice President of Data and AI, Oladimeji (Dimeji) Farri has to say about the rapidly evolving landscape of agentic AI patient engagement and what healthcare leaders should keep in mind as they determine how best to adopt and integrate this technology into their organizations.
Capabilities and Functionality
1. How do agentic AI engagement solutions advance hyper-personalized care?
We’re at an exciting moment where advanced technology makes truly personalized patient engagement a reality. With a next generation platform for precision care and deep industry expertise, Get Well empowers patients to navigate the often complex healthcare landscape with greater confidence and ease. By ensuring each patient’s unique needs and preferences are understood and addressed, we help connect them with care that genuinely fits — all with the aim of driving better health outcomes. Generative and agent-based AI take this even further by enabling us to build AI assistants that draw meaningful insights from patient data, align with trusted clinical guidelines and care plans, and deliver timely, context-aware support tailored to each individual’s situation. This not only ensures patients receive the right care at the right time, but also helps reduce the administrative burden on busy care teams, freeing them to focus more on what matters most: caring for patients.
2. How should an AI assistant handle complex patient queries or situations that require human clinical judgment?
AI assistants should be designed to enhance — not replace — the clinical care team, by reasoning through extensive clinical context and knowledge to support timely, accurate decision-making and more efficient healthcare operations. When faced with complex patient scenarios, well-configured AI assistants can operate within clearly defined guardrails and follow robust escalation pathways, ensuring that all tasks and recommendations remain ethical, responsible and clinically trustworthy — with a human expert always available for oversight.
Beyond guardrails and escalation, implementing observability to detect performance drifts after deployment is rapidly becoming a best practice in modern Large Language Model Operations (LLMOps) and the emerging AgenticOps landscape. Ultimately, to deliver hyper-personalized patient engagement safely and effectively, it is crucial to build in clear mechanisms for escalation to qualified clinical professionals and to continuously monitor deployed AI solutions. This enables healthcare organizations to make informed decisions about when to refine models, strengthen guardrails and ensure that AI-driven care remains aligned with the highest standards of clinical excellence and patient trust.
3. What safeguards and escalation protocols should a solution have in place when the AI cannot provide an adequate response or identifies a potential issue?
I firmly believe that the very first safeguard for any AI initiative should be a thoughtful evaluation of whether the problem truly requires a generative and agentic AI solution. Misapplying AI where it adds unnecessary complexity often results in wasted investment and misplaced frustration with the technology itself. Once it’s clear that the product requirements and solution space genuinely warrant a generative or agentic approach, the next priority should be to design robust guardrails and escalation pathways. These mechanisms should activate whenever the system approaches its defined limits in knowledge, performance or acceptable error thresholds. Equally important is ongoing oversight: the AI’s compliance with these safeguards must be actively monitored throughout the solution’s lifecycle. As AI continues to evolve rapidly, I expect that fully automated monitoring and corrective actions for performance drift and policy adherence will soon become as standard as DevOps practices and agile development are today.
4. How will agentic AI enhance clinical workflows, reduce cognitive burden and improve patient outcomes?
Clinical workflows today are bogged down by mountains of paperwork exchanged among patients, clinicians and payors. Imagine our AI assistants as tireless backstage partners, seamlessly handling routine tasks — automating documentation during bedside rounds, generating summaries and referrals, and even populating forms on demand. Beyond paperwork, these assistants can analyze patient records to spot care gaps, then proactively schedule appointments and send reminders for screenings or follow-up visits to bridge those gaps. They can also deliver tailored education and interventions — think personalized medication reminders or lifestyle tips — to boost adherence and outcomes. By taking on these repetitive yet critical duties, the AI assistants can free up healthcare teams to focus on what matters most and help curb avoidable costs tied to readmissions and complications.
Architecture and Data
5. How does agentic AI integrate with existing EMR systems and workflows?
At Get Well, our next-generation platform is designed with a robust, secure data layer that seamlessly connects to EHR systems. Using SMART on FHIR along with other trusted healthcare messaging standards and technologies, we extract and integrate the most relevant patient information to build a more complete picture of each individual’s health journey. Beyond data integration, we harness both descriptive and predictive analytics to augment our AI agents and assistants. This enables them to access longitudinal, comprehensive patient synopses, which supports more personalized patient interactions and meaningful outcomes analysis.
By doing so, we help our clients drive engagement while advancing their clinical, operational and financial goals. In a nutshell, we don’t just deliver technology — we provide actionable intelligence that supports better care experiences and measurable results.
6. How should an AI patient engagement platform be built to ensure scalability, performance and reliability for a variety of environments, including large health systems?
Since I’ve already highlighted various aspects of our platform in earlier responses, I’d like to expand specifically on how we ensure scalability, reliability and performance across our ecosystem. As part of the SAI Group, we leverage strong synergies with our sister companies — for example, RhythmXAI — to shape a cohesive and forward-thinking approach to our cloud infrastructure, technology stack and operational workflows. Our strategy integrates best practices in DataOps, DevOps, LLMOps, Infrastructure-as-Code, and the use of AI-powered coding assistants. By thoughtfully combining these critical components, we’re able to deliver generative and agenting AI solutions that are not only innovative but also robust, resilient, and adaptable to the diverse needs of modern healthcare systems. Our commitment to this foundation ensures that the solutions we bring to market remain state-of-the-art and ready to evolve alongside our customers’ requirements.
7. Could you elaborate on the data strategy and architecture underlying agentic AI, specifically how patient data is managed securely to support AI functionalities like longitudinal data analytics?
In addition to leveraging descriptive and predictive analytics to power our AI agents, we take a thoughtful and pragmatic approach to assessing the level of customization and investment required to meet each customer’s specific goals and measurable outcomes. While training or fine-tuning AI models with customer data is not always necessary given the diverse range of use cases we support, we do have the capability to deliver this when it adds clear value. When customization is required, we implement robust data anonymization measures and optimize the use of cost-effective compute resources, ensuring that custom AI models and agents are developed and deployed efficiently and securely. More broadly, our AI assistants are built on a foundation of best-in-class open-source and proprietary models, orchestrated through agentic workflows within our platform. This modular approach allows us to adapt our solutions to a wide variety of customer requirements while maintaining scalability, performance and cost-effectiveness.
Implementation and Future Vision
8. What is the typical implementation process for integrating Get Well’s advanced agentic AI solution, Opal, into an environment, and what strategies do you employ to ensure successful adoption?
Opal should be seen as the friendly, customer-facing personality of our AI assistants. Integrating Opal successfully hinges on two key factors: the deployment of the underlying AI capabilities and thoughtful workflow integration that supports meaningful engagement, both within the hospital and beyond. The integration process typically starts with a comprehensive review of the hospital’s existing workflows and data infrastructure to ensure Opal can connect seamlessly and work in harmony with current systems. Subsequently, Opal is configured to align with each hospital’s unique care pathways, discharge processes and communication protocols. For patients, our goal is to provide an effortless, always-available assistant that goes beyond reminders — it proactively educates, supports and motivates them to follow through with critical health tasks. This element of encouragement is vital: Opal doesn’t just nudge patients but gives them compelling reasons to stay on track. For clinical teams, we focus on demonstrating clear value: Opal is designed to lighten their workload, free up valuable time and enable them to dedicate more attention to direct patient care and meaningful interactions.
9. How do you anticipate AI platforms and tools like Opal will evolve to meet the changing needs of the industry and health systems?
At Get Well, we’re committed to accelerating innovation in AI-powered patient engagement. Our solutions are evolving to be even more responsive to each patient’s unique context — taking into account not only clinical needs but also social, emotional and lifestyle factors. Our goal is to become the trusted, lifelong AI advocate for patients, supporting healthcare systems across the U.S. and around the world. Looking more broadly, as health systems continue to expand their digital ecosystems, intelligent agents and AI assistants will play an increasingly vital role. They’ll integrate effortlessly with EHRs, remote monitoring devices, and other digital tools, helping to orchestrate a more connected, seamless experience for both patients and care teams.
10. How is Get Well’s approach to agentic AI different than what others are doing, like EHRs and point solutions?
Unlike many solutions that focus solely on either clinical workflows or in-room patient interfaces, Get Well’s generative and agentic AI approach focuses on being a true patient advocate throughout the entire care journey. We go beyond task automation by personalizing engagement in a way that considers each patient’s medical history, social needs, and emotional well-being — whenever and wherever. Rather than simply delivering information, our AI solutions aim to build trusted relationships and encourage positive behavior change. The result? Better outcomes, a more meaningful patient experience, and a deeper sense of human connection (augmented by AI) at every step.
Ready to bring agentic AI to your hospital?
Talk to our team today to schedule a personalized demo of Opal and discover how proactive AI engagement can streamline workflows, enhance patient education, and improve post-discharge outcomes.
