Using generative AI to help dog owners make smarter health decisions

If you’re one of the 65 million US households with a dog, this might sound familiar: your dog’s behavior suddenly changes, and you don’t know whether to worry. Maybe your dog refuses to go up or down the stairs, lies down in the middle of a walk, or stops eating. If it’s a weekend or holiday, then you’re unsure whether to pay for an emergency vet visit—or perhaps the nearest vet is hours away.

What if you could get trusted answers from a leading veterinary college, any time of day, from anywhere? That’s the idea behind Big Red Bark Chat, a generative AI-powered chatbot developed by the Cornell Richard P. Riney Canine Health Center (RCHC) at the Cornell University College of Veterinary Medicine, in collaboration with Amazon Web Services (AWS).

Big Red Bark Chat enables pet owners to get on-demand access to canine health information curated by RCHC faculty. The solution was built using several AWS services in a seamless architecture designed for reliability, scalability, and security.

Advancing canine health through research and engagement and outreach

Cornell’s College of Veterinary Medicine has long been a pioneer in canine health, developing more than half the vaccines veterinarians use today. In 2020 and 2021, donor support helped launch the RCHC at the College of Veterinary Medicine, with a mission of expanding public education and accelerating research around canine health.

Today, RCHC funds Cornell-based canine health research and distributes roughly $1 million annually across studies in infectious disease, cancer, aging, genetics, genomics, and more. Along with supporting research, RCHC provides up-to-date canine health information to the public. Staff members field public inquiries through telephone calls, emails, and the website that are often related to common canine health concerns, behavioral issues, and more in depth canine health diseases.

But demand quickly outpaced capacity in the new center—especially after a pandemic-era boom in first-time dog owners.

“There aren’t enough behaviorists or specialists in the country to service all of the questions that come in,” said Dr. Rory Todhunter, director of the RCHC. And, when dog owners are anxious about a new symptom or unexpected behavior, he added, “They want immediate answers from trusted sources, but might not be able to reach their veterinarian right away. We needed a solution that could bridge this critical information gap while emphasizing the importance of proper veterinary care.”

Building a generative AI solution with AWS

AWS has supported Cornell’s College of Veterinary Medicine for more than a decade, enabling a close working relationship. When the RCHC team began exploring AI-powered tools, their AWS account team connected them with the AWS Generative Artificial Intelligence Innovation Center (GenAIIC)—a group of strategists, data scientists, and engineers who help users build practical generative AI applications.

Within three months, the AWS GenAIIC team created a proof-of-concept chatbot that responded to common canine health questions using health articles from the Riney Canine Health Center, Cornell’s proprietary knowledge base, Consultant, and faculty lecture notes and presentations.

Later dubbed Big Red Bark Chat (a nod to Cornell’s “Big Red” nickname), the tool allows owners to evaluate symptoms, understand potential next steps, and decode veterinary medical reports—while reinforcing that it does not replace a veterinarian.

“Big Red Bark Chat is transforming how we share our expertise with the public,” said Dr. Todhunter. “We’re now able to provide trusted information to dog owners at any time of day, helping them make better decisions about when to seek veterinary care.”

Designing a secure and scalable generative AI architecture

At the heart of the solution is Amazon Bedrock, the fully managed service for foundation models (FMs) at AWS. The application uses Anthropic’s Claude Sonnet 4 as its primary model for handling user interactions, providing nuanced responses to complex canine health questions.

To make sure responses are grounded in accurate veterinary information, the team implemented Amazon Bedrock Knowledge Bases to ingest and index Cornell’s extensive collection of canine health articles, research papers, and educational materials. This approach ensures the model’s responses are rooted in Cornell’s authoritative content, rather than potentially outdated or inaccurate information.

The solution was containerized using Docker and deployed on Amazon Elastic Kubernetes Service (Amazon EKS) for scalability, with an Application Load Balancer (ALB) handling traffic distribution. The infrastructure is managed through Terraform, allowing for consistent deployments and updates.

Security was paramount in the design, with strict access controls, content moderation through custom prompts to filter inappropriate requests, and clear disclaimers emphasizing the tool’s informational—not diagnostic—nature.

Figure 1. Big Red Bark Chat architecture

Making veterinary expertise accessible to all dog owners

Because the data sources for Big Red Bark Chat are technical (for example, research papers and educational materials used to teach veterinarians), the team spent a lot of time making the tool more accessible to the average pet owner. The chatbot provides answers at a high school reading level and offers sample prompts and tips for asking questions that elicit the best answers.

As generative AI necessitates human refinement and feedback, RCHC faculty submitted questions and evaluated the answers to check and confirm accuracy. After a period of fine-tuning, Big Red Bark Chat made its debut in March 2025 with a few hundred students and faculty. Early feedback has been largely positive and is the resource has now undergone a broader public release.

“This is just the beginning of our journey to make veterinary knowledge more accessible,” noted Dr. Todhunter. “As we continue to refine this tool, we see tremendous potential for supporting not just pet owners, but also veterinary students, practitioners in remote areas, and animal welfare organizations worldwide.”

Democratizing access to veterinary expertise

Cornell’s collaboration with AWS shows how generative AI can be thoughtfully applied to democratize specialized knowledge while maintaining appropriate professional boundaries. Big Red Bark Chat gives dog owners direct access to Cornell-vetted content, reinforcing the university’s mission to advance animal health through research, education, and outreach.

The project serves as a model for other educational institutions seeking to expand public access to scholarly research through responsible AI implementation. Combining institutional expertise with cutting-edge generative AI technology allows Cornell to create new pathways for knowledge to flow from the university to the broader public—ultimately benefiting dogs and the people who care for them.

To learn more about how AWS can help your institution build, deploy, and scale AI solutions, contact the AWS Public Sector Team.

Read related stories on the AWS Public Sector blog:

• Seven lessons learned: Driving AWS Cloud adoption in research at Saint Louis University
• Cloud Innovation Center launch at University of Pittsburgh
• How the University of Johannesburg uses chatbots to increase student engagement and reduce costs
• Cal Poly seeks to transform procurement process with generative AI-powered scope of work generation