Why AI matters at this scale

The May Institute is a large nonprofit organization providing educational, rehabilitative, and behavioral healthcare services to individuals with autism spectrum disorder (ASD), intellectual and developmental disabilities, brain injury, and other special needs. Founded in 1955 and operating across multiple states, it employs between 1,001 and 5,000 staff, offering a continuum of care including schools, adult services, and behavioral health clinics. Its mission centers on evidence-based practice to improve client independence and quality of life.

For an organization of this size and mission complexity, AI presents a critical lever to enhance both operational efficiency and clinical outcomes. The human services sector is notoriously resource-constrained, with thin margins and heavy reliance on staff time. At a scale of thousands of employees and clients, small efficiency gains in scheduling, documentation, or personalized intervention planning can compound into significant financial and qualitative benefits, freeing resources for direct care. Furthermore, the institute's longitudinal work with clients generates vast amounts of behavioral and progress data, which, if harnessed responsibly, can unlock insights for more proactive and personalized support.

Concrete AI Opportunities with ROI

1. Predictive Staffing and Incident Prevention: By applying machine learning to historical data on client behaviors, environmental factors, and staff interventions, the institute could build models to forecast periods of high client acuity or potential behavioral incidents. The ROI is twofold: optimized staffing reduces overtime and burnout, while preventive interventions decrease costly crisis responses and improve client stability, potentially leading to better outcomes and funding compliance.

2. Clinical Documentation Automation: Clinicians and direct support professionals spend excessive time on mandatory documentation. Natural Language Processing (NLP) tools can transcribe session notes, auto-populate standardized forms, and flag inconsistencies. This directly translates to reclaimed billable hours, reduced administrative overhead, and more consistent data for reporting and research, offering a clear, quantifiable return on investment.

3. Personalized Program Adaptation: AI algorithms can analyze individual client progress across therapeutic and educational activities to recommend adjustments to their treatment or learning plans. This moves the model from standardized protocols to dynamically personalized care, which can accelerate skill acquisition and improve goal attainment rates. The ROI manifests in more effective service delivery, potentially leading to higher client retention and satisfaction, which are key for reputation and funding in a competitive nonprofit landscape.

Deployment Risks for a 1001-5000 Employee Organization

Implementing AI at this scale carries specific risks. First, change management across dozens of locations and diverse roles (from clinicians to administrators) is monumental. A top-down mandate will fail without deep engagement and training. Second, data silos and quality are a major hurdle. Client data may be fragmented across different legacy systems, schools, and geographic programs, requiring significant upfront investment in data integration and governance before models can be trained reliably. Third, ethical and regulatory scrutiny is intense. Any AI tool affecting client care must be transparent, auditable for bias, and compliant with HIPAA and other disability-rights regulations. A perceived misstep could damage trust irreparably. Finally, sustained funding for AI pilots is challenging in a nonprofit model where budgets are tight and donor funds are often restricted to direct service. Projects must be phased to show quick, tangible wins to secure ongoing investment.