How to Choose an AI Keynote Speaker: Seven Questions Every Event Organiser Should Ask

Choosing an AI keynote speaker starts with your audience.

What do they already understand? What are they uncertain about? What should they be able to discuss or do differently after the session?

A board briefing, a sales conference and an engineering event need different approaches. The subject may be AI, but the purpose of the session should be much more specific.

As someone who delivers AI keynotes and leadership workshops, I believe organisers should ask for evidence before making a booking. That includes evidence of subject knowledge, speaking ability and relevance to the people in the room.

Here are seven questions I would use to assess any speaker, including me.

1. Can they explain what your audience will take away?

“An inspiring talk about the future of AI” leaves a lot open to interpretation.

Ask the speaker to describe three things your audience should leave understanding.

For a leadership team, those might be:

• How to identify a worthwhile first AI project.
• Where human approval is needed when AI takes actions.
• How to measure whether a change improves the work.

For a customer-facing team, the priorities could be preparing for conversations, finding reliable information and checking AI-generated responses.

The takeaways should match your event’s purpose. They should also be realistic for the time available.

Ask: “What will our audience leave better equipped to do?”

2. What evidence sits behind their examples?

An impressive result needs context.

If a speaker describes an organisation saving time or increasing revenue, ask where the information came from. Was it measured by the client? Reported by a technology supplier? Observed during a demonstration? Or estimated for an illustrative example?

Each can be useful, provided the distinction is clear.

My work through AI Ireland and the AI Awards gives me access to examples of organisations applying AI. In my AI Awards leadership article, I use published finalist stories to explore questions about people, processes, platforms, data and services.

Those stories provide a starting point for discussion. Being shortlisted is not independent verification of every performance claim, and a result in one organisation is not a promise that another will achieve the same.

That distinction matters when your audience includes people who will make investment decisions.

Ask: “Which claims can we verify, and what limitations should our audience understand?”

3. Can you watch them explain something properly?

A showreel can give you a sense of energy and stage presence. A longer speaking excerpt helps you assess clarity.

Watch how the speaker introduces an unfamiliar idea. Can you follow the explanation without already knowing the terminology? Do the examples help? Does the speaker connect the technology to a decision the audience recognises?

If possible, watch a question-and-answer exchange. It can reveal whether the speaker listens carefully, handles uncertainty and responds to the question actually asked.

You can review my speaking highlights alongside the audience and event context.

Ask: “Can we see a speaking example relevant to our audience?”

4. How will they tailor the session?

Personalisation should affect the substance of the talk.

A useful briefing should cover the audience’s roles, existing AI experience, business priorities and concerns. It should also establish what tools people can use and what information must remain confidential.

For example, a demonstration that assumes unrestricted access to company data may be unsuitable for a regulated organisation. A detailed technical explanation may be valuable for engineers but need a different treatment for a mixed conference audience.

My published client case studies describe different contexts, including engineers at Medtronic and small business owners at LEO Kilkenny. The relevant question for an organiser is how the approach fits their own room.

Ask: “What will you change after learning about our audience?”

5. Do the demonstrations show judgement as well as capability?

A demonstration should help people understand both an opportunity and the decisions involved in using it.

Suppose an AI system produces a customer email. The useful discussion includes the information it used, the assumptions it made and what a person should check.

If an agent can send the email, another question appears: who gave it that authority?

A strong demonstration makes those boundaries visible. It should also be clear whether the demonstration is live, recorded or based on a prepared example.

For event planning, agree practical details in advance: connectivity, display requirements, account access and a fallback if the tool is unavailable. Use approved demonstration data.

Ask: “What will this demonstration teach our audience to check?”

6. What does client feedback actually demonstrate?

Look for feedback that tells you something specific.

Did the speaker make a difficult subject understandable? Adapt to the audience? Encourage useful discussion? Give participants something they could apply?

Those observations help you assess the session.

They do not automatically establish a financial return or a lasting change in behaviour. An enthusiastic response on the day and a documented business outcome three months later are different forms of evidence.

For example, the Medtronic feedback on my case-study page describes a team feeling encouraged to explore new ways of working. That is useful audience feedback. It should not be presented as a measured productivity result.

Where appropriate, ask for a reference from an organiser with a similar audience.

Ask: “What can a previous client tell us about the preparation, delivery and relevance?”

7. Is a keynote the right format for the outcome?

A keynote can give a large audience a shared understanding and a clear set of questions to take back to work.

A workshop provides more room to apply those ideas: examine a process, test an approach, compare priorities or decide what to do next.

If you want a leadership team to leave with an agreed shortlist of AI opportunities, allow time for participation and discussion. If you want a conference audience to understand a changing subject and see practical possibilities, a keynote may be the right starting point.

You may need both, but each session should have a distinct purpose.

Explore the differences between my AI keynotes and AI leadership workshops.

Ask: “Does this format give us enough time to achieve the outcome we want?”

A short checklist for your speaker brief

Before contacting speakers, write down:

• Audience: Who will attend, and how familiar are they with AI?
• Purpose: What should change as a result of the session?
• Context: Which business problems or decisions matter most?
• Evidence: What speaking footage, references or examples do you need?
• Format: How much time is available for the talk, demonstrations and questions?
• Boundaries: What information, tools or topics require care?
• Follow-through: Who will help the audience act on what they learn?

Send the same brief to each speaker you are considering. Compare their responses against your requirements.

The most useful question is whether they can make AI relevant, understandable and actionable for your audience—and show you evidence that supports that judgement.

Planning an AI keynote for your event?

If you are organising a conference, leadership offsite or company event, tell me about your audience and the questions you want the session to address.

Check my availability and discuss your event.