GW RhythmX launches to unify clinical, financial, payer, and social data for precision care
Get Well and RhythmX AI are merging under SymphonyAI to form GW RhythmX, a new AI-driven personalized care platform. The goal: surface actionable insights that improve operations, outcomes, and margins across health systems.
RhythmX AI founder Deepthi Bathina will serve as CEO. Michael O'Neil, founder and CEO of Get Well, will become vice chair. Together, the companies serve 150 health systems and 85 million patients, including eight million veterans.
Why this matters for healthcare leaders
GW RhythmX plans to bring data from more than 10 sources into a single platform-combining EHR data, clinical policies and guidelines, financial and payer data, and social determinants. The platform is EHR-agnostic and built to give clinicians precise next steps while helping executives manage throughput, length of stay, and revenue integrity.
- Clinical: Evidence-aligned recommendations and earlier identification of risk.
- Operational: Discharge orchestration, resource allocation, and throughput optimization.
- Financial: Coding and billing support to reduce leakage and improve yield.
- Patient experience: More relevant engagement at every point in the care journey.
What the tech already delivers
Since launching in 2023, RhythmX AI has generated more than 300,000 uniquely personalized treatment recommendations, optimized coding and billing, and used analytics to predict disease progression and surface likely undiagnosed conditions.
Get Well reports its customers cut hospital readmissions by more than 65%. Earlier this year, it launched Opal-an AI patient assistant built on RhythmX technology-to automate post-discharge check-ins, track real-time discharge progress, improve care plan adherence, and resolve barriers like transportation. Health systems using these capabilities include Adventist Healthcare, Ascension, UMass Memorial, BayCare, and LifeBridge Health.
Leadership perspective
"The combined company brings the needed resources, technology and scale together to enable health systems to thrive in the new era of patient-centric, AI-powered precision care," said Dr. Romesh Wadhwani, SymphonyAI Group chairman.
"With our companies coming together as one, we have the opportunity to forever change the way medicine is practiced," said CEO Deepthi Bathina. "By orchestrating all of healthcare's data into one platform of actionable intelligence, we're helping clinicians focus on what matters most, giving patients a more connected experience and enabling health systems to achieve new levels of operational and financial performance."
What to do now
- Map your data landscape: confirm access to EHR, claims, clinical policies/guidelines, and SDOH data; define priority use cases.
- Align governance: set standards for data quality, consent, PHI security, and model oversight with clinical, IT, and compliance stakeholders.
- Target quick wins: start with readmissions, discharge coordination, revenue cycle coding support, and risk stratification.
- Plan integration: validate EHR-agnostic connectors, interface throughput, and workflow fit for clinicians and care managers.
- Measure impact: track KPIs like LOS, readmits, denials, coding accuracy, and patient-reported experience.
- Upskill teams: train clinical ops, revenue cycle, and care management on AI-assisted workflows and exception handling.
What to watch
- Integration timeline and go-live readiness across multiple EHRs.
- Model transparency, bias mitigation, and human-in-the-loop guardrails.
- Payer-provider collaboration on prior auth, risk adjustment, and value-based care.
- Demonstrable ROI within 6-12 months on operational and financial metrics.
Learn more about the portfolio and partnerships at SymphonyAI. For context on clinical AI in documentation and ambient tools, see Microsoft's healthcare AI work with Nuance's Dragon solutions here.
If your team is building practical AI skills for clinical operations and revenue cycle, explore role-based training at Complete AI Training.
Your membership also unlocks: