Photo: Cecilie Arcurs/Getty Images
Newly announced artificial intelligence applications highlight the shift toward domain-specific automation, where reasoning and native integration aim to improve efficacy and safety.
Three recent product announcements of new artificial intelligence tools show how AI is evolving across healthcare use cases and hint at where it could be headed next.
- Ambience Healthcare this month launched a new Chart Chat for Nursing, a generative artificial intelligence feature that gives nurses the ability to query electronic health records systems at the point of care.
- Corti has introduced AI-driven medical coding built on a proprietary model that the company says has outperformed competing genAI coding tools built on OpenAI's ChatGPT and Anthropic's Claude.
- Ensemble and Cohere, meanwhile, have teamed up to build a revenue cycle management-native large language model (LLM) that they say can help reduce provider administrative burdens.
Ambient genAI retrieves EHR data for nurses
Ambience said it has expanded its Chart Chat tool to allow nurses to talk to Epic Systems electronic health records and ask questions and get answers during inpatient conversations.
It was built specifically for use on the hospital floor to quickly obtain patient medication histories, lab trends and more, the company said in an announcement on April 1. The tool is intended to help nurses understand diagnoses and care pathways or access general clinical information at the point of care.
"Chart Chat for Nursing meets nurses where they already are, inside the EHR, and gives them the full picture of every patient in seconds," explained Nikhil Buduma, Ambience cofounder and CEO, in a statement.
Results are generated as text inside the Ambience module inside the EHR, a spokesperson told Healthcare IT News on Tuesday. While Chart Chat is currently compatible with Epic EHRs, the company said it is exploring integrations with additional systems.
All of the AI's responses are governed by a three-tier safety architecture, including evaluations during deployments, quality monitoring in real time and nurses' feedback, according to the company.
Multi-agent medical coding built on new ML model
Corti's new medical coding product, called Symphony for Medical Coding, offers a multi-agent workflow that examines clinical text and analyzes it against coding rules, the company said in its April 1 website announcement.
It runs a four-stage agentic reasoning workflow on every request to focus on the active diagnosis, first filtering out details from notes that do not need coding, including historic conditions that are no longer being managed.
The ML model behind it, named Code Like Humans, was trained on 5.8 million electronic health records from 1.8 million patients, Corti said.
For each identified diagnosis, the model queries the ICD-10 alphabetical index to locate the relevant terms and all associated sub-entries against quality standards and then generates a full candidate code set – just as a trained coder would.
The model returns primary code along with ranked alternatives along with source text that triggered the prediction and justifications for auditing the results.
When evaluated across five datasets spanning healthcare settings in the U.S. and UK, the LLM outperformed OpenAI and Anthropic models by more than 25%, the company said.
"Most AI systems fall short in medical coding because they treat it as labeling, not reasoning," said Lars Maaløe, cofounder and CTO of Corti. "Correct coding depends on evidence, context, hierarchy and guideline interpretation."
Symphony is available as an API endpoint, as a Model Context Protocol, and through enterprise and cloud deployments, the company said.
RCM-native LLM for healthcare ops
While most large language models fall short on regulatory nuances, payer needs and processes foundational to healthcare operations, Ensemble, an RCM company, and Cohere, an AI company, said that the two will build a fully custom model that relies on Ensemble's operational experience and data.
"Our associates' operational knowledge, well‑defined processes and insight into payer behavior are shaping how models learn and solve problems that will help us reduce friction across the patient financial journey and continue delivering the trusted results health systems count on," said Judson Ivy, Ensemble's president and CEO, in an announcement on March 31.
"By pairing Ensemble's deep domain expertise with our secure, enterprise‑grade AI capabilities, we can create agents that deliver greater accuracy, consistency, and reliability while meeting the highest standards of privacy and security," added Aidan Gomez, Cohere's CEO.
The model will be fine-tuned on RCM tasks and embedded into AI agents that power health system operations from patient intake to account resolution, the companies said. It will not be trained on identifiable client data or protected health information, they added.
Andrea Fox is senior editor of Healthcare IT News.
Email: afox@himss.org
Healthcare IT News is a HIMSS Media publication.
