"The biggest benefit [of working with the Microsoft AI Co-Innovation Labs] was being able to rapidly reach POC and begin testing. We were able to quickly design with engineers from the Lab team, create infrastructure for efficient data cleaning, and coordinate seamlessly to build something out that would have taken our team ~3 months in a matter of weeks." – Annalina Che, CEO, Stemtology
...

Co-Innovation Challenge

Regenerative medicine is a quickly evolving field that promises therapies for previously incurable diseases and helps advance medical research. Stemtology is an innovator in this critical space, accelerating effective disease treatments with AI-driven technologies. Their AI platform integrates large language models (LLMs) and graph neural networks (GNNs) to hypothesize, simulate, and optimize treatment plans for inflammatory, immune, and degenerative diseases.

stemtology_case_study_healthcare_worker_reviewing_medication

Stemtology’s work addresses known limitations in their industry, like slow experiment times, siloed research publishing, and overwhelming data complexity. Even as they address these challenges, the Stemtology team needed to partner with the “best of the best”1 to build a scalable prototype and get backend support on a solution in Azure.

They partnered with the Microsoft AI Co-Innovation Labs to build a generative AI solution that could find and analyze medical research data and generate treatments for osteoarthritis , with a plan to apply this solution across other diseases later.

In the Lab

Our AI Co-Innovation Lab team in San Francisco brought our extensive experience in building gen AI prototypes using Microsoft Azure and AI technologies to collaborate on a multi-agent research and development solution.

The Lab team leveraged Azure OpenAI GPT and Azure Cognitive Search for data processing, hypothesis generation, and validation. They also used their extensive knowledge of cutting-edge Microsoft technologies to use the brand-new Azure AI Agent Service, empowering Stemtology with a more powerful and scalable agentic AI.

Stemtology Agents Workflow
Stemtology Agents Workflow

The first set of agents act as research assistants, searching for medical studies from the PubMed database and pulling that data for insights and reports. The next agent takes that research and generates osteoarthritis treatment plans based on disease-specific research studies that the science team can then analyze and test.

This solution integrates directly into Stemtology’s business model, acting as a research and development engine and working in tandem with scientists in the lab team to accelerate design, testing, and optimization of synthetic cell therapies.

Solution Impact

Stemtology is already seeing the benefits of their solution with their science team, including significant operational cost savings and reduced time spent on labor-intensive tasks like reading and analyzing research papers. As a result, scientists can shift their focus to higher-level design tasks and explore other disease treatments.

In deploying the solution, Stemtology expects to achieve even greater outcomes: 

  • Cut experimental timelines by up to 50%.
  • Achieve ≥90% predictive accuracy for treatment outcomes, ensuring feasible and effective therapeutic solutions.
  • Scale the solution to manage over 100 diseases.
  • Improve regenerative outcomes for each disease treated.
stemtology_case_study_doctor_and_patient_talking

This advanced application of gen AI positions Stemtology as a leader in regenerative medicine, offering transformative benefits for healthcare providers, researchers, and patients by accelerating drug discovery and improving access to personalized treatments.

By collaborating with our team of experts in the AI Co-Innovation Labs, Stemtology can realize and scale this solution with the power of agentic AI in Azure.

1 Quoted from Annalina Che, CEO, Stemtology.

立即加入

微软人工智能和物联网实验室

即刻申请，获得与技术专家共同设计解决方案的机会。