Dallas-based Pieces Technologies, a healthcare generative AI company for care teams, announced it incorporated Amazon Web Services' genAI offerings to create Sculpted AI, a technology that can tailor to health systems' specifications and embeds AI within electronic health record clinical workflows. 

Sculpted AI leverages Amazon Bedrock, Amazon Elastic Compute Cloud and Amazon SageMaker Canvas to draft inpatient clinical summaries in real-world clinical environments autonomously. 

The platform also helps providers identify patient discharge barriers and predict discharge dates, auto-generate snapshots of a patient's status, identify social determinants of health and detect critical care deterioration. 

Sculpted AI can be tailored to the unit, specialty or service level using a "health system's change management principles with a high degree of human oversight and in rapid fashion."

Pieces said it uses adversarial and collaborative AI, human-in-the-loop systems control and auto-prompt engineering, combining clinical expertise and technology to promote iteration and safety. 

"There is a major opportunity and need to harness generative AI at the point of care to derive better insights and reduce administrative burden on clinicians," Dan Sheeran, general manager of healthcare and life sciences at AWS, said in a statement. 

"In particular, being able to integrate generative AI directly into EHR clinical workflows via Amazon Bedrock and Pieces' technology can provide meaningful value," Sheeran said. "Pieces' continued advancements demonstrate its culture of rapid innovation and responsible approach to AI for the benefit of both patients and clinicians."

THE LARGER TREND

Amazon announced the general availability of Bedrock, aimed at helping organizations build and scale generative AI applications, in September. The service includes Meta's Llama 2 through a managed API, one of several foundation models trained on large amounts of data. 

The company also announced the availability of Amazon Titan, Amazon Quicksight and the future release of its AI-powered coding companion, Amazon CodeWhisperer.