LAS VEGAS, Nev.— At the 2025 HLTH conference in Las Vegas, GE HealthCare unveiled health system partnerships and the latest research projects that are part of its AI Innovation Lab.
The company is working with the Queen’s Health Systems in Honolulu and Duke Health in Durham to advance the development of its new AI-driven hospital operations software, which will become part of CareIntellect. CareIntellect, a generative AI platform, is a hub for various GE HealthCare applications and was first announced at HLTH last year. The idea is to enable health systems to easily deploy new applications without a product-by-product integration approach.
“Now you can really have a common data layer, that’s all the data is structured in a way that you can analyze and plug into. You can bring more applications, whether that’s on the operations side, on the care delivery side,” Taha Kass-Hout, GE HealthCare’s global chief science and tech officer, told Fierce Healthcare.
CareIntellect’s first application was intended for oncology. Last week, CareIntellect added a Perinatal product, developed alongside HCA Healthcare clinicians. It aggregates data from mothers and infants, including vitals, to provide real-time analytics for clinicians to review patient status and search events without time-consuming manual review.
Queen’s Health Systems is the largest health system in Hawaii and is the only Level 1 trauma center in the Pacific Basin. Accepting transfers is crucial to maintaining care access, but the health system has had issues with throughput and length of stay, per executives. They had to deny about half of all transfer patients as a result. The health system engaged with GE HealthCare to address these issues, deploying its Command Center solution, which helps health system administrators manage patient flow, streamline operations and make data-driven decisions in real-time.
With the software, the health system has seen a 22% jump in patient transfer admissions in 10 months. It can now accept 100 more monthly transfer patients, crucial to a health system that is on an island state. They have also seen a 41% decrease in ED length of stay, saving an estimated $20 million on an average reduced length of stay of a day in the first year. Command Center, used by 500 organizations today, will become an application part of CareIntellect in the future.
While Queen’s has had an established IT steering committee, “the conversation hasn’t been near as exciting as it is currently and not as passionate as it is currently,” Ashley Shearer, VP of care coordination, patient flow, geriatrics and inpatient rehab at Queen’s, told a room of reporters at the HLTH conference Monday morning. The health system is cautiously excited about the possibilities of AI in unlocking the full potential of this data. “The more data you unlock, the better, cleaner, more targeted your workflows become.”
CareIntellect is cloud-first, powered by Amazon Web Services. Since it has zero footprint for setup on-site, this streamlines deployment at organizations. Executives say they can deploy CareIntellect within days, which makes the offering more financially viable for healthcare organizations.
All organizations interested in CareIntellect have one thing in common, per Kass-Hout: they want to make their data actionable. “A common theme across all health systems is: help me translate all this data to action,” he said. “If you don’t solve for that, it’s going to be very hard to bring more innovation into the health system.”
At the HLTH conference, GE HealthCare also announced the latest research projects that are part of its AI Innovation Lab, launched last year. The initiative aims to accelerate early-concept AI innovations within the company. Interested GE HealthCare customers can support some of the research efforts. The latest projects include agentic AI for radiology and testing GE HealthCare’s MRI research foundation model with two academic medical centers.
GE HealthCare is attempting to develop what it says is the first agentic AI diagnostic imaging assistant that would be integrated into devices, specifically designed for radiology. It is based on a model trained on over 200,000 MRI images. Mass General Brigham and the University of Wisconsin-Madison will assess the model across operational and clinical use cases, paired with their own data and using machine learning to test and refine the technology.
Other AI Innovation Lab projects include evaluating incidental findings in CT scans with agentic AI; developing less energy-intensive tomographic imaging; and supporting engineers who repair hospital equipment with a conversational bot that can recommend fixes while citing sources.
GE HealthCare is committed to carefully testing and validating its AI models, executives say. It has already published a peer-reviewed publication, with another forthcoming and a third oral abstract, that explain its algorithms and how they work -- reviewed alongside clinicians.
In 2026, Kass-Hout predicts that there will be more reimbursement pathways for AI solutions. There has been a big change in regulations compared to past approaches, he said to reporters at the conference. Examples include efforts like FDA’s recent predetermined change control plans guidance and the 92229 CMS code, put out in 2021. Key to reimbursement will be proving high-quality systems, Kass-Hout said: “Outcome-based is going to become very, very key to these reimbursements.”