Ambience Healthcare rolled out an ICD-10 clinical document integrity assistant for inpatient care, aiming to help hospitalists document more accurate clinical notes and reduce billing and coding issues downstream.
Built on top of OpenAI GPT-5’s advanced reasoning capabilities with Ambience’s proprietary models, the platform was designed to address the gap between clinical complexity and clinical documentation integrity, according to Ambience executives.
Ambience says its inpatient CDI platform was designed specifically for hospital medicine and engineered for compliance, auditability and clinical nuance.
"The inpatient setting is where the most complex patients are treated and where documentation is most burdensome,” Mike Ng, president and chairman of Ambience Healthcare, said. “By addressing CDI upstream, at the point of care, we’re helping hospitals create records that are accurate, audit-ready, and reflective of true clinical complexity.”
The company already built out an ICD-10 assistant for ambulatory care. Ambience designed its latest artificial-intelligence-powered CDI assistant to address the complexities of inpatient workflows and coding, noted William Morris, M.D., Ambience's chief medical officer and a hospitalist by practice.
"These are highly complex patients that are admitted to the hospital where it's not just you, it's a team sport. Sometimes these patients show up and they're completely undifferentiated. They just show up with an admitting diagnosis. It's very time compressed, you have to do everything during that hospitalization to understand that patient," Morris said.
Clinicians in the inpatient setting face nonstop documentation demands across admission, rounding, handoffs and discharge. Many hospitals have CDI programs and coding professionals to improve the accuracy and completeness of documentation, but they often work retrospectively, long after clinicians have cared for the patients.
Accurate and comprehensive diagnosis documentation is critical for patient safety, quality outcomes and reimbursement, Morris noted. Any issues downstream can increase audit risk and reimbursement delays.
Ambience says its new inpatient CDI solution eliminates this friction by equipping hospitalists with real-time support for diagnostic specificity, POA designations and complication tracking, embedded in hospitals' existing Epic workflows.
The goal is to streamline processes, reduce follow-up queries and ensure high-quality, cost-effective care, executives said.
Ray Chen, M.D., head of clinical AI at Ambience Healthcare, said the technology represents a "fundamental shift in how clinical documentation supports both care delivery and compliance."
“By enabling accurate coding and diagnostic specificity at the point of care, we reduce administrative burden and ensure that the clinical story is faithfully captured—which translates into more informed treatment decisions, better continuity of care, and ultimately improved outcomes for patients," Chen said.
Founded in 2020, Ambience built a platform that uses AI for documentation, clinical documentation integrity and point-of-care coding. The AI documentation software records patient appointments, automates documentation with ambient listening and preps them with specialty-specific chart summaries.
More than 40 U.S. health systems—including Cleveland Clinic, UCSF Health, Houston Methodist and Memorial Hermann—have deployed its ambient AI technology, the company said.
Ambience integrates directly into the electronic health record and supports more than 100 specialties, including complex and underserved domains like oncology, psychiatry and emergency medicine. The company attests that its platform adapts to the unique context of each care setting and specialty without requiring workflow redesign or staff retraining.
In July, the company secured $243 million in series C funding, marking one of the largest health tech raises so far in 2025. The funding, backed by Oak HC/FT, a16z, the OpenAI Startup Fund and Optum Ventures, among other investors, boosted the company's valuation to $1.25 billion.
A key feature of the inpatient CDI solution is an explainable audit trail as every suggested ICD-10 code includes clinical rationale and supporting documentation references. The solution also has built-in guardrails to ensure suggestions meet Centers for Medicare & Medicaid Services guidelines and audit requirements.
Ambience works closely with OpenAI to access its frontier AI models and then, using its own proprietary models, works to fine-tune and train the technology for healthcare-specific applications.
"This is highly tuned, task-specific trained models. Trained to think like a hospitalist on admission. Think like a world-class coder and quality abstractor. We have access to the core frontier capabilities, and then we just take it to the next level in terms of fine tuning, training and reinforcement learning within the domain of healthcare, because that's our core spot. We're not building foundation models. We want to take world-class leading ones and then make them purpose-built to do very complex tasks," Morris said.
Ambience sees opportunities to provide what Morris refers to as "documentation decision support" by using reasoning models to surface insights in the clinical data.
"What gets me really excited with these new frontier capabilities is reasoning over data. What is really tough for clinicians is not what I said. But what should I have said in that office? What are the things that I don't know that I don't know?" Morris said. "Reasoning over clinical data is not just about time efficiency, it's actually better quality of care, it's safer care, and it's more affordable care."
He added, "Areas that we're excited about, our ability to search the chat, and we partner with authoritative content companies that provide that anchoring backbone of here's the evidence, and what are you doing around it? In the workflow, as I'm documenting and thinking about the patient, it feels like a way more normal and organic opportunity to inject nudges and assist the clinician."
As it continues to build out new AI tools, Ambience also sees opportunities to support allied health and nursing staff with AI assistive agents.