Technology & Innovation

Strategic Planning for AI Deployment: A Collaborative Approach

AI is rapidly reshaping medical imaging. Attendees of the Radiological Society of North America (RSNA) Annual Meeting have witnessed the shift firsthand, as AI vendors have expanded from a limited presence to a dominant feature of the exhibit floor.

This growth is reflected in regulatory trends, too. FDA-cleared AI-enabled medical devices have increased from approximately 700 in 2023 to more than 1,000 in recent years, with the majority approved within a short timeframe. As adoption accelerates, imaging leaders are challenged to move beyond awareness and develop structured approaches to evaluate and deploy these technologies effectively.

Moving From Awareness to Strategy

Following a recent RSNA meeting, our imaging leadership team identified the need to transition from general awareness of AI to a defined system-wide strategy. Key questions included how to evaluate the growing number of vendors, how to align AI initiatives with organizational priorities, and how to ensure the right stakeholders were engaged.

To address these challenges, we established a system-wide Imaging AI Steering Committee to guide evaluation, selection, and deployment efforts.

Stakeholder Alignment and Governance

The steering committee included radiologists, IT leadership, the chief medical information officer (CMIO), PACS administrators and analysts, and radiology administrative leaders. Each stakeholder contributed essential expertise, including clinical workflow, technical infrastructure, and operational performance.

This multidisciplinary approach ensured that decisions were aligned with enterprise priorities and reduced the risk of fragmented or siloed implementation.

Defining Clinical and Operational Priorities

A critical early step was defining the problems AI was intended to solve. Rather than pursuing technology for its own sake, the committee focused on areas where measurable impact could be achieved.

Radiologist workflow emerged as a primary priority. Improving efficiency and case prioritization has direct implications for report turnaround times, patient throughput, and clinical outcomes — particularly in time-sensitive scenarios.

At the same time, IT stakeholders evaluated key technical considerations, including cloud architecture, cybersecurity, and system integration. These factors are essential to ensuring scalability and long-term success.

Platform Versus Point Solution Strategy

One of the most important decisions involved selecting between individual point solutions and a platform-based approach. A point solutions approach usually involves a single AI algorithm designed for a specific purpose whereas a platform-based approach allows multiple algorithms to be hosted on a single enterprise platform. The committee elected to pursue a platform strategy based on several advantages:

  • Simplified deployment for IT teams managing competing priorities
  • Scalability, allowing additional algorithms to be implemented over time
  • Single-vendor accountability, reducing the complexity and challenge of coordination
  • Supports a more sustainable, enterprise-wide strategy for AI integration

Vendor Evaluation and Selection Process

To guide vendor selection, the committee developed a standardized evaluation framework aligned with organizational priorities. Core criteria included:

  • Workflow integration with PACS and EHR systems
  • Enterprise scalability across multiple sites
  • Financial return on investment, including operational efficiency and downstream revenue
  • Data security and regulatory compliance
  • Continuing go-live support and performance monitoring

A formal request for information was issued to a select group of vendors. Responses were consolidated into a comparative matrix to facilitate objective evaluation. The committee then narrowed the field to a small number of finalists that participated in on-site presentations prior to final selection.

Implementation and Ongoing Governance

Following vendor selection, the project transitioned to IT leadership for contract negotiation, implementation planning, and go-live execution. However, the role of the steering committee did not end at implementation.

Ongoing governance remains critical to monitor system performance, evaluate clinical impact, and address challenges such as AI model drift or workflow integration issues. AI drift is the gradual decline of an AI or machine learning model’s performance over time. It occurs because the real-world environment, user behavior, or incoming data changes, rendering the model's original training data outdated.  Workflow integration refers more to the user experience and ensuring that the models continue to support the radiologist’s workflow. This continued oversight and engagement ensures that AI solutions deliver sustained value over time.

Practical Takeaways for Radiology Leaders

  • Establish governance early: A formal steering committee provides structure and alignment.
  • Engage multidisciplinary stakeholders: Clinical, technical, and operational perspectives are all essential.
  • Focus on workflow impact: Solutions should improve efficiency without disrupting existing processes.
  • Think enterprise-wide: Scalability is critical for long-term success.
  • Prioritize measurable value: Evaluate both clinical outcomes and financial impact.

A Plan for Long-Term Success

AI adoption in medical imaging continues to accelerate, creating both opportunity and complexity for health systems. Successful deployment requires more than evaluating individual technologies — it demands a structured, collaborative approach.

By establishing multidisciplinary governance, defining clear priorities, and applying a rigorous evaluation process, radiology leaders can move from experimentation to scalable implementation. This approach not only improves decision-making but also creates shared ownership and supports long-term success.