I’ve sat in a lot of hospital board meetings discussing AI. Some go well. Many don’t.

The ones that don’t typically share a pattern: the CIO presents AI as a technology initiative, gets caught up in technical details, and loses the room. Board members nod politely while thinking about the agenda item that matters more to them.

If you’re preparing to present an AI strategy to your board, here’s what I’ve learned about what works.

Know Your Audience

Hospital boards include clinicians, community representatives, financial experts, and governance professionals. Their AI understanding varies wildly. The clinician might know exactly what diagnostic AI can and can’t do. The community representative might have read one alarming article about AI bias.

Don’t pitch to the person with the most AI knowledge. Pitch to the person with the least, then be ready to go deeper if asked.

This means:

• Avoid technical jargon entirely
• Use analogies that connect to governance concepts they understand
• Focus on outcomes and risks, not technology features
• Be prepared for questions that seem basic (they’re not testing you; they genuinely don’t know)

Frame It as a Business Decision, Not a Technology Decision

The first question board members should be able to answer after your presentation: “Why are we doing this?”

If the answer is “because AI is the future,” you haven’t made the case.

Better answers:

• “To reduce diagnostic errors in emergency radiology by 15%”
• “To address waitlist growth without proportional increases in specialist FTE”
• “To improve patient safety by flagging medication interactions we’re currently missing”
• “To meet TGA requirements that will affect our existing clinical systems”

Each of these is concrete, measurable, and connected to outcomes the board already cares about.

Structure Your Presentation Around Their Questions

After dozens of board presentations, I can predict the questions. Build your presentation to answer them preemptively:

“How much will this cost?”

Total cost of ownership over five years, not just implementation. Include:

• Licensing fees (often per-study or per-patient)
• Infrastructure (hardware, cloud, network upgrades)
• Integration (typically higher than vendors estimate)
• Training and change management
• Ongoing governance and monitoring

“What are the risks?”

Be honest about risks. Boards respect CIOs who don’t pretend everything will go perfectly. Key risks to address:

• Patient safety risks if AI makes errors
• Regulatory compliance risks
• Cybersecurity and data privacy risks
• Vendor sustainability risks (will this company exist in five years?)
• Workforce risks (clinician resistance, skill gaps)

“How will we know it’s working?”

Define success criteria before implementation. What metrics will you track? What performance levels constitute success? At what point would you consider discontinuation?

“Who else is doing this?”

Boards like knowing they’re not first movers (too risky) or last movers (falling behind). Reference implementations at comparable organisations—similar size, similar patient population, similar service profile.

“What’s the timeline?”

Be realistic. AI implementations typically take 50-100% longer than initially estimated. Build contingency into your timeline.

The One-Page Summary

Before your detailed presentation, create a one-page summary that a board member could read in two minutes. Include:

• What you’re proposing (one sentence)
• Why it matters (two sentences)
• Total five-year cost
• Key risks and mitigations (bullet points)
• Success metrics
• Recommended decision

Board papers are long. Attention spans are finite. Give them the key points upfront.

What Not to Do

Don’t demo the technology. A demo seems like a good idea. It rarely is. Demos break at the worst moments. They raise questions about features that aren’t relevant to the decision. They shift focus from outcomes to technology.

Don’t overwhelm with data. Two or three compelling statistics are better than twenty. Pick the ones that matter most.

Don’t downplay regulatory complexity. If TGA approval is required, say so clearly. If ADHA guidelines apply, explain what that means. Boards appreciate understanding the regulatory landscape.

Don’t promise what you can’t deliver. Overpromising on AI outcomes is endemic in our industry. The board will remember what you said when results come in. Underpromise and overdeliver.

Don’t treat it as a one-time decision. Frame AI as a program, not a project. The initial decision is just the beginning. The board should expect to see updates as implementation progresses.

Addressing the Hard Questions

Sometimes board members ask questions that don’t have comfortable answers. Prepare for these:

“Is this going to replace clinical staff?”

Be honest. In most cases, current AI augments rather than replaces. But workforce implications exist. Productivity improvements might reduce future hiring. Some roles might change significantly. Pretending there are no workforce impacts undermines your credibility.

“What happens if the AI makes a mistake and a patient is harmed?”

Explain your clinical governance framework. Acknowledge that AI isn’t perfect. Describe how you’ll detect errors and respond. Make clear that clinical responsibility remains with clinicians, not AI systems.

“What does this mean for our liability?”

You probably can’t answer this fully without legal advice. If you haven’t engaged your legal team on AI liability, do that before the board presentation.

“How do we know the AI isn’t biased?”

Explain what bias testing the vendor has done, and what bias monitoring you’ll implement. If you can’t answer this confidently, it’s a gap in your due diligence.

After the Presentation

Assuming you get approval, establish a reporting cadence. Quarterly updates to the board during implementation. Annual reviews once the system is operational.

Each update should return to the success metrics you defined. Are you on track? If not, why not? What are you doing about it?

Boards don’t like surprises. Regular, honest updates maintain confidence even when things don’t go exactly as planned.

The Bigger Picture

Board engagement on AI isn’t just about getting approval for a specific initiative. It’s about building organisational capability for the AI-enabled future.

The board needs to develop governance literacy about AI. Each presentation, each update, each discussion builds that literacy. Over time, the board becomes a more effective governance body for AI initiatives.

That’s worth investing in—even when it makes individual presentations take longer.

---

Dr. Rebecca Liu is a health informatics specialist and former Chief Clinical Information Officer. She advises healthcare organisations on clinical AI strategy and implementation.