Clinical Scorecard: Rethinking the ROI of AI

At a Glance

Key Highlights

• The article argues that AI’s value can include improved decision quality and stronger human work, not only reduced costs or time.
• Practice leaders can use AI to clarify problems, test assumptions, explore outcomes, and develop counterarguments before making consequential decisions.
• Mapping revenue-cycle handoffs with AI revealed reliance on individual knowledge and gaps in documented standard operating procedures.
• A custom-coded cataract quoting tool calculated patient responsibility from team selections, reduced complexity, and increased patient concierge staff confidence in educating patients.
• The article cautions against over-automation that sacrifices human judgment or the experiences people need to grow.

Guideline-Based Recommendations

Diagnosis

Management

• Give AI the problem, context, and constraints, then ask it to interview you one question at a time to clarify the issue.
• Ask AI to challenge assumptions, identify potential outcomes, and develop counterarguments.
• Use AI to document workflows and examine handoffs, ownership, and gaps in standard operating procedures.
• Consider systematizing complex calculations when human involvement does not improve the output, while preserving staff attention for patient communication and connection.
• Use leadership meetings to discuss how teams apply AI to test assumptions, explore alternatives, and improve decision-making.

Monitoring & Follow-up

• Review workflow handoffs and clarify ownership where responsibilities are assumed rather than explicit.
• Maintain a living training resource and improve accountability across the workflow.

Risks

• AI can reinforce assumptions embedded in the questions it is given.
• The article cautions against over-automation that sacrifices human judgment and the experiences people need to grow.

Patient & Prescribing Data

No patient treatment population or prescribing data are provided.

The article describes an AI-supported cataract surgery quoting tool for estimating patient responsibility; it does not provide clinical treatment recommendations.

Clinical Best Practices

• Use AI to challenge and pressure-test decisions rather than relying on it only for confirmation.
• Redesign work around what technology makes possible instead of simply adding AI to existing processes.
• Systematize repetitive or logic-based work where appropriate, while retaining human judgment and patient connection where they matter.
• Document workflows and make process ownership explicit.

Related Resources & Content