5 ways Amazon Pharmacy is using generative AI to improve healthcare

5 ways Amazon Pharmacy is using generative AI to improve healthcare

Amazon Pharmacy Generative AI is being used to dispense prescriptions more quickly and accurately, making customer service faster and more helpful, and ensuring customers are stocked with the correct amount of medication. Amazon Pharmacy is also working on developing artificial intelligence applications to provide more transparent pricing for prescription drugs, allowing customers to purchase them at the best prices.

“Today, Amazon is leveraging generative artificial intelligence to deliver a better pharmacy experience for our clinical teams and customers,” said Alexandre Alves, senior principal engineer, Amazon Pharmacy. “We see an opportunity to rethink the flow of information, and that’s both exciting And transformative.”

Amazon Pharmacy is a full-service pharmacy in the Amazon.com store. Customers can use Amazon Pharmacy to purchase doctor-prescribed medications and have them conveniently delivered to their door. Prime members enjoy free two-day shipping and a variety of savings, including the Prime Prescription Savings benefit and RxPass, which gives you unlimited access to eligible prescription drugs for just $5 a month. All Amazon Pharmacy customers can contact a pharmacist 24/7 to discuss questions about their medications.

Last year, Amazon Pharmacy doubled the number of customers it served, and artificial intelligence is helping the team scale efficiently. To enhance their generative AI capabilities, Alves and his team used multiple pre-trained models from Amazon Bedrock and Amazon SageMaker. “By building our solutions on Amazon Web Services, our teams are able to move faster and focus on creating the best experience for our customers,” Alves said.

Increase processing speed and get medications to customers faster

Amazon Pharmacy Prescription Bottles.

Alves has spent his entire career as a computer scientist using data and deep learning models to help streamline supply chains and predict customer demand. Now applying his expertise to healthcare, he says there is some truth to the cliché about doctors having bad handwriting – although these days, in most cases prescriptions are typed and passed through the clinical team Delivered by electronic transcription service.

However, digital prescriptions can still contain confusing or inconsistent language. “The data has no standard structure. For example, in a field labeled ‘Direction,’ we might find phrases like ‘taken by mouth’ or ‘orally,'” he said. “By making these directions more structured, we can expand Operational scale, dispensing medications faster, more efficiently and, importantly, reducing human error. “

Amazon solves this challenge by generating artificial intelligence models to run on raw “unstructured” data. The model uses a process called named entity recognition, using categories such as “dose” and “frequency” to create text structure. From there, it helps Amazon Pharmacy clinical staff fill prescriptions and provide clear instructions to patients.

Each prescription is still reviewed by a pharmacist because the provider’s guidance may be incorrect for reasons that cannot be detected by AI. For example, sometimes prescribers will put “volume” in a field that means “strength.” But by combining generative AI methods with the expertise of pharmacists, Amazon was able to speed up order processing by 90% and, most importantly, reduce human error rates.

Stock up on the right medications

Photo by Amazon pharmacist Andrew Vo.

Leveraging his supply chain efforts, Alves and his team also use artificial intelligence and machine learning (ML) to forecast demand. This helps Amazon stock the right medications at each location so they are ready to dispense when prescriptions arrive.

“Generative AI helps us synthesize data to experiment with different inventory scenarios. This can help us improve our ability to predict the drugs needed at each location,” Alves said. “Similarly, we have a variety of jars used to dispense medications, and AI can help us better determine which jar is best for each medication.”

Help customers find the best prices

Amazon pharmacy displayed on mobile phone screen.

Amazon Pharmacy is using machine learning and generative artificial intelligence along with large language models (LLMs) to provide real-time insurance estimates for prescription drugs without requiring customers to enter their insurance details. Customers can shop for the best prices on their medications by reviewing the list of pricing options on the Amazon Pharmacy website, which includes estimated insurance pricing and ways to save with the Prime Prescription Savings benefit, receive pass or Automatic coupons.

“Viewing drug prices directly on a pharmacy website or mobile app is surprisingly novel,” Alves said. “The LLM can create immediate value for clients by providing up-to-date pricing, enabling better decision-making.”

Provide better support to customers

Image of a person holding a mobile phone and an Amazon pharmacy prescription bottle.

Generative AI can help Amazon Pharmacy clinical and customer service teams answer questions faster. It can review internal documentation pages and knowledge bases and summarize them in a useful and clinically appropriate way. Amazon also enhances the process by applying context to guardrails for each issue.

“Of course, in the field of pharmacy, accuracy and safety are extremely important. We used another model to verify the answer,” Alves said. “This model is independent of other models and trained based on feedback from our staff. Our clinical and customer service representatives review everything before speaking with customers.”

Deliver drugs more efficiently

Amazon pharmacy prescription bottles in fulfillment center.

Alves says there’s always a balance between executing tasks now and waiting until the work can be done in batches. “By improving our ability to ingest data and interpret context, generative AI can help us improve these predictions and bulk decisions and help customers get their prescriptions faster,” he said. “If you don’t need a refill within a month, the request may wait until we can process it in bulk. If you need the medication urgently, we will move it to the front of the line.”

Alves added: “We continue to discover new uses for generated artificial intelligence. We are excited to not only further improve the Amazon Pharmacy customer experience, but also significantly enhance the role of our pharmacy employees to help create better health care outcomes. “

Learn more about Amazon Pharmacy.

Leave a Comment

Your email address will not be published. Required fields are marked *