Reclaiming time for patient care using Amazon Bedrock with Plume IA

For many clinicians, patient care doesn’t end when an appointment is over. After the final patient leaves, more work awaits: completing notes, referrals, forms, and other documentation. Some even wake early each morning to finish paperwork before returning to the clinic, balancing patient care with an administrative workload. “As doctors, we’re trained to see patients and take care of people, not to make notes,” says Dr. Jasmin Landry, cofounder and CEO of Plume IA.

Existing AI medical scribes were often designed for English-language use and healthcare workflows outside Quebec. Quebec clinicians, however, often switch between French and English during consultations while following local documentation practices.

Dr. Landry first tested a prototype from his living room, even carrying his computer to the clinic. He then partnered with software engineer Marc-André Paquin to develop the initial Plume IA solution. The two-person founding team chose AWS to support sensitive healthcare data and provide access to foundation models as generative AI technology evolved.

Plume IA used Amazon Bedrock, the solution for building generative AI applications, as the foundation. After receiving patient consent, clinicians record consultations through Plume IA. Using AI models in Amazon Bedrock, the recorded conversations are converted into structured clinical notes, referral letters, insurance documentation, and other healthcare records. Clinicians review the generated documentation before adding it to the patient’s record, which reduces administrative work while maintaining oversight of clinical information. Users can also create customizable templates that reflect their preferred documentation formats across different healthcare professions.

Using Amazon Bedrock gives Plume IA access to a choice of foundation models, so the team can evaluate models for different processes as capabilities and costs change while building a secure application trusted with sensitive healthcare data. Because Amazon Bedrock presents models within the same security framework, Plume IA does not need to reassess its approach each time it tests a model. “This is where AWS capabilities stand out,” says Dr. Landry. “It really helped us build our product quickly with the security needed for healthcare data.”

To host and train proprietary AI models and automatically scale compute resources based on demand, Plume IA also uses Amazon Elastic Compute Cloud (Amazon EC2), which offers secure and resizable compute capacity. The company adjusts compute capacity as demand changes, increasing it during periods of higher use and reducing it when demand is lower. Running workloads in the AWS Canada (Central) Region helps the company meet Canadian data residency requirements, supporting the privacy expectations of healthcare organizations.

A geriatric medicine team increased weekly patient evaluations from 5 to 12–15 after adopting Plume IA. When Dr. Landry was practicing, he saved about 1–2 hours each day in the clinic and could see three to five additional patients per day in emergency medicine. According to Canada Health Infoway, clinicians using AI scribes like Plume IA report up to 70 percent reduced administrative burden, and more than 80 percent said they felt more engaged with patients during visits.

Beyond improving productivity, Plume IA is helping clinicians reclaim personal time. “A colleague of mine was waking up every morning at six to finish his paperwork from the night before,” says Dr. Landry. “With Plume IA, his life changed. He has more time for his family and himself. He can breathe now.”

Plume IA has grown largely through recommendations among healthcare professionals across Quebec and has reached 5,500 paying users. The company is prototyping patient engagement workflows using Amazon Connect, an AI-powered cloud contact center, to support future patient triage. Plume IA also plans to expand into other healthcare settings, markets, and languages. “There’s a shortage of healthcare professionals,” says Dr. Landry. “Our solution reduces pressure on the system, helping clinicians see more patients and offering better care.”