AI Agent Operational Lift for Sanford Research in Sioux Falls, South Dakota

Why AI matters at this scale

Sanford Research, based in Sioux Falls, South Dakota, is the research arm of Sanford Health, one of the largest rural health systems in the U.S. With 201–500 employees, it conducts biomedical, clinical, and translational research, focusing on cancer, genomics, pediatrics, and rare diseases. Its integration with a health system provides a unique asset: access to longitudinal patient data from electronic health records (EHRs). This mid-sized institute operates at a scale where AI can be transformative—large enough to have meaningful datasets but small enough to avoid the bureaucratic inertia of massive academic medical centers.

For research organizations of this size, AI addresses critical bottlenecks: manual data extraction, slow patient recruitment for trials, and the overwhelming volume of scientific literature. By adopting AI, Sanford Research can amplify its research output, compete more effectively for NIH grants, and translate findings into clinical practice faster. The institute’s existing data assets, combined with modern cloud-based AI tools, create a fertile ground for high-ROI projects.

Three concrete AI opportunities

1. Accelerating clinical trial recruitment with NLP Manually screening EHRs for trial eligibility is labor-intensive and error-prone. Deploying natural language processing (NLP) to parse unstructured clinical notes can automatically identify candidate patients, reducing screening time by up to 70%. ROI: faster trial enrollment, lower per-patient costs, and increased grant success rates due to efficient feasibility demonstrations.

2. Genomic data integration for precision medicine Sanford Research generates multi-omics data but often struggles to integrate it with clinical phenotypes. Machine learning models can fuse genomic, proteomic, and clinical data to identify novel biomarkers. ROI: higher-impact publications, intellectual property generation, and potential licensing revenue from diagnostic tools.

3. Digital pathology with deep learning Pathology slides are a rich, underutilized data source. Training convolutional neural networks to detect and grade tumors can augment researchers’ capabilities, enabling large-scale retrospective studies. ROI: reduced pathologist time, faster turnaround for research projects, and new avenues for computational pathology grants.

Deployment risks for a mid-sized institute

Despite the promise, Sanford Research faces risks typical of its size band. Data privacy and HIPAA compliance are paramount; any AI system handling patient data must be rigorously secured. The institute may lack in-house AI engineering talent, requiring investment in training or partnerships with universities. Model bias is another concern—training on a predominantly rural, Midwestern population may limit generalizability. Finally, change management can be challenging: researchers accustomed to traditional methods may resist AI-driven workflows. Starting with low-risk, high-visibility pilots and securing executive sponsorship from Sanford Health will be key to overcoming these hurdles.