Sayvant saves 50,000 hours of emergency clinician charting time with Microsoft Azure

Sayvant is using generative AI to improve clinical charting, save time, and produce higher-quality documentation.

Using Microsoft Azure solution Azure OpenAI Service, Sayvant built an AI solution that generates acute care charts to comply with data security standards for the healthcare industry and align with documentation requirements.

Sayvant has saved 50,000 hours of clinician time, as the AI solution transcribes and generates personalized discharge instructions in more than 30 languages and helps clinicians focus more on face-to-face patient care instead of manual documentation.

Acute care clinicians are often overwhelmed by the demands of documentation, reducing the time they can spend with patients. Sayvant, a health technology start-up, is on a mission to change that. By providing AI-driven clinical documentation solutions for emergency and other acute care providers, Sayvant is harnessing the power of AI and large language models to enhance the quality of patient care and improve financial and operational efficiency.

Empowering clinicians to spend more time with patients

Delivering high-quality care in acute care settings is challenging. Clinicians are burdened with extensive documentation required for transitions of care, reimbursement, and quality measurement. Sayvant, codesigned and codeveloped with Vituity, one of the largest democratically physician-owned healthcare delivery organizations, recognized this pressing issue and wanted to make a difference. Such a nuanced workflow meant the team had to create a product specifically tailored for an acute care setting.

Between 2017 and 2021, denial rates by payors surged by 20%,¹ and Centers for Medicare & Medicaid Services coding changes in 2023 have placed the burden on clinicians to provide extensive documentation regarding their medical decision-making.² This unfortunately means less face-to-face time with patients and higher burnout rates, resulting in staffing shortages and inefficiencies across hospital systems. On average, emergency medicine clinicians spend over two hours per shift, or about 20% of their time, on charting alone.³ “Every patient going to the emergency room is having their worst day,” says Dr. Andrew Napier, Cofounder of Sayvant. “If you can make that slightly better, whether it’s a faster visit, a more efficient visit, or an improved care outcome, ultimately, that’s why we’ve built Sayvant.”

“Every patient going to the emergency room is having their worst day. If you can make that slightly better, whether it’s a faster visit, a more efficient visit, or an improved care outcome, ultimately, that’s why we’ve built Sayvant.”

Dr. Andrew Napier, Cofounder, Sayvant

Sayvant envisioned a generative AI solution that would transcribe, summarize, and generate patient charts so that clinicians could have more face-to-face interactions while maintaining document quality. The company had an existing relationship with Microsoft and chose to build its solution on Microsoft Azure offering Azure OpenAI Service. “We can do everything from accessing pay-as-you-go services for Azure OpenAI models to renting graphics processing units (GPUs) to self-host our own models,” says Justin Mardjuki, CEO of Sayvant. “This allows us flexibility to customize and build quickly.”

Sayvant built a multitenant architecture to ensure scalability, security, and compliance to meet demanding regulatory requirements like SOC 2 and the Health Insurance Portability and Accountability Act. “We’re building the highest-performing, secure clinical documentation solution that meets the demands of acute care settings,” says Mardjuki. “Microsoft is the preferred cloud vendor for a healthcare enterprise leader, and choosing Azure for Sayvant was a natural decision.”

“We’re building the highest-performing, secure clinical documentation solution that meets the demands of acute care settings. Microsoft is the preferred cloud vendor for a healthcare enterprise leader, and choosing Azure for Sayvant was a natural decision.”

Justin Mardjuki, CEO, Sayvant

Transforming patient care by decreasing documentation burden in more than 30 languages

Collaboration with clinicians is at the heart of Sayvant’s research and development. The company built its solution based on extensive feedback from hundreds of clinicians working at sites ranging from rural urgent care centers to level 1 urban trauma centers. “Given the amount of protected health information we’d be handling in these institutions, we wanted to ensure that they felt comfortable with our choices regarding infrastructure, scalability, security, and privacy,” says Mardjuki. “We’ve been very happy with that decision.”

Sayvant transcribes conversations between clinicians and patients in near real time and in more than 30 languages. The solution then creates robust, clinically defensible charts for healthcare providers to review, along with personalized discharge instructions. “There’s a pretty radical transformation when clinicians walk into the room with Sayvant in their pocket,” says Mardjuki. “The doctor doesn’t have to worry about getting every word right because Sayvant helps them generate that chart. They can have a face-to-face conversation with the patient.”

Reducing discharge delays by 40%

Sayvant adoption has spread to nearly 70 sites since its private beta app launched in the summer of 2024. To date, more than 30,000 shifts have been completed using Sayvant, and charting time for emergency providers has been significantly reduced, from 10 minutes to less than 90 seconds per patient, saving an estimated 50,000 hours.

Extra one-on-one time with patients and clear discharge instructions has led to a 40% reduction in discharge delays at one site, reducing patient time in the waiting room and in the emergency room. “As labs and images are returned, clinicians can have ongoing conversations with Sayvant about diagnoses,” says Mardjuki. “This allows them to focus on the right next step of care for the patient instead of the right next step in documentation.”

Some clinicians now use Sayvant exclusively, an adoption rate 10 times higher than other solutions on the market. “I have been using Sayvant exclusively for the last three months and can never go back to the original ways of charting,” says Dr. Cameron Nouri, Medical Director of the Department of Emergency Medicine at the Community Hospital of San Bernardino in California. “I am able to quickly and efficiently see more patients and finish my charts before getting out on time.”

“I have been using Sayvant exclusively for the last three months and can never go back to the original ways of charting. I am able to quickly and efficiently see more patients and finish my charts before getting out on time.”

Dr. Cameron Nouri, Medical Director of the Department of Emergency Medicine, Community Hospital of San Bernardino

Expanding the impact of AI on healthcare

Sayvant is excited to partner with more healthcare providers and clinicians, and it plans to continue investment in its best-of-breed large language model infrastructure. “We’re building Sayvant with input from experts in the field so that we can really validate the impact that we are having,” says Mardjuki. “As we expand the number of sites, clinicians, and patients that we support every day, we’re excited about deepening our AI investments to ensure that we continue to improve the speed, accuracy, and quality of our acute care documentation output.”

Discover more about Sayvant on LinkedIn.

Medical Device Disclaimer: Microsoft products/services (1) are not designed, intended or made available as a medical device(s), and (2) are not designed or intended to be a substitute for professional medical advice, diagnosis, treatment, or judgment and should not be used to replace or as a substitute for professional medical advice, diagnosis, treatment, or judgment.

1 “More than 30% of hospitals are near the ‘danger zone’ of denial rates,” Jeff Legasse, Healthcare Finance, June 2021.
2 “CPT® Evaluation and Management (E/M) code and guideline changes,” American Medical Association, January 2023.
3 “Electronic health record documentation times among emergency medicine trainees,” Scott Crawford, Igor Kushner, Radosveta Wells, and Stormy Monks, Perspectives in Health Information Management 16, January 2019.