A technology conference director at a financial services firm described it this way: the speaker had an impressive video reel, a recognizable roster of past clients, and a confident pitch call. Everything looked right until the day before the keynote, when the speaker's team emailed requesting a dressing room with specific lighting, a catering order not on the rider, and a 90-minute AV setup window in a venue that had allocated them 30. The keynote itself was generic. The Q&A exposed gaps. None of it should have been a surprise, because the warning signs were there weeks earlier.
Booking an AI speaker is not like booking other speakers. The category is crowded with people who have genuine expertise in adjacent fields but limited ability to deliver for a specific audience, alongside speakers whose authority rests entirely on media presence rather than real-world implementation. The field also moves fast enough that content sharp eighteen months ago may read as outdated to anyone in the room paying attention.
The red flags below are specific, observable, and consistently predictive of problems. Learning to catch them before you sign is one of the most valuable skills an event planner can develop.
The Video Reel Is Not Evidence
Every speaker sends a highlight reel. The better question is whether you can find an unedited, full-length recording of a past presentation.
Highlight clips tell you almost nothing. They show the moments a speaker or their team selected as their best work, trimmed to remove dead air, pacing problems, and weak transitions. A 90-second clip of a speaker landing a strong line at a major conference doesn't tell you how they handled the 35 minutes before it, whether the audience was engaged or restless, or what happened when a skeptical executive challenged their claims during Q&A.
What you actually want is a full recording from a comparable audience, ideally from a corporate event rather than a staged speaker showcase. Watch how they handle transitions between sections. See whether they check in with the audience or barrel through their material. If the speaker or their team will only provide curated clips, treat that as information in itself.
Also worth scrutinizing: the client logo wall. A slide deck of 40 company logos sounds impressive until you ask follow-up questions. When were they there? What was the event? How large was the audience? Logos from three or four years back in a field that has changed this dramatically tell you very little about whether this speaker can serve your audience today.
What the Discovery Call Reveals (and What Skipping It Signals)
A speaker who asks about your audience before talking about themselves is a meaningful positive signal. A speaker whose team tries to close the booking without a discovery call is a red flag.
The structure of the call reveals intent. A speaker doing legitimate preparation will want to know who is in the room, what their current relationship with AI looks like, what strategic context is driving this event, and what you want the audience to leave believing or doing differently. These questions cannot be answered by reading your event website.
Speakers who skip this step, or who treat the discovery call as a sales presentation rather than a research conversation, are usually delivering a lightly adjusted version of the same talk to every audience. Your room will feel like it is watching content designed for someone else.
Red flags to listen for during the call:
- The speaker or their manager does most of the talking
- Questions about your audience feel perfunctory ("so what industry are you in?") rather than substantive
- The speaker claims they can address any aspect of AI for any audience level (genuine expertise has edges)
- They don't ask whether a previous speaker has covered similar content with this group (experienced speakers know audiences dislike repetition)
- They can't describe what they would not cover or what falls outside their domain
For events where the audience includes practitioners such as engineers, data teams, or technical product managers, ask directly how the speaker adjusts content for a room with deep AI knowledge. Speakers who work primarily with C-suite audiences often struggle to hold rooms that can actually pressure-test their claims.
If you're building a framework for pre-booking evaluation, our AI keynote speaker guide includes a question set designed for exactly this stage of vetting.
"I Can Speak About Anything AI" Is a Warning Sign
Real expertise has a shape. A speaker who has spent years on AI in healthcare operations knows the specific regulatory frameworks, clinical workflow integration challenges, and data governance issues that keep hospital CIOs up at night. That depth is what makes them worth the fee. It is also what makes them the wrong choice for a retail audience trying to understand AI in personalization.
The red flag is a speaker who presents as fully interchangeable across AI topics and industries. Claims like "I tailor everything to your industry" or "I've spoken to audiences in every sector" aren't differentiators. They signal that the speaker's value proposition is delivery style, not domain knowledge.
This matters more now because audiences are more sophisticated than they were two or three years ago. An HR leadership conference that has already sat through multiple AI sessions doesn't need an overview of what large language models are. They want someone who has worked with organizations at their stage and can speak specifically to their context.
When evaluating whether a speaker's expertise applies to your event, look at their last eight to ten speaking engagements, not their career history. If those recent bookings span wildly different industries without a connecting thread, dig into what is actually consistent about the content. Depth in a domain matters more than breadth of appearances.
For a breakdown of how AI expertise maps to specific industry audiences, the AI strategy topics section outlines the content areas where speaker specialization carries the most weight.
Riders, Technical Requirements, and the Last-Minute Demand
A speaker's technical rider should arrive before the contract is signed. If it arrives after, or if the speaker's team says "we'll send that over once we're confirmed," that is a problem.
The rider tells you what kind of professional you're dealing with. A speaker who does live AI demonstrations will have specific requirements written down because they have learned from experience what happens without them: bandwidth minimums, backup slide formats, dual display requirements, contingency plans for demo failures. These details should not emerge as a surprise negotiation 72 hours before the event.
When reviewing a technical rider, watch for two opposite problems.
No rider at all, particularly for a speaker promising live AI demos, suggests they either lack experience with what happens when technical requirements aren't met, or they are planning to deliver a simpler presentation than the one they pitched.
An extremely demanding rider with requirements that standard conference venues cannot reasonably support is also worth scrutinizing. Not because the requirements are wrong, but because a speaker with genuine conference experience knows what venues can typically accommodate and will flag incompatibilities proactively.
The question to ask directly: "What is your backup plan if internet connectivity is insufficient for your demonstration?" A speaker who has a real answer (a locally run model, pre-recorded fallbacks, a static version of the same content) is a speaker who has done this before. A speaker who looks surprised by the question probably hasn't.
Contract and Bureau Terms That Signal Risk
Most event planners focus on the fee when reviewing a speaker proposal. The fee is usually the least informative indicator of whether the engagement will go well.
What is worth reading carefully:
Cancellation terms: Standard speaker contracts use a tiered structure, typically no penalty beyond a set number of days out, graduated fees as the event date approaches, and full fee forfeiture within 30 days. A contract weighted dramatically against you regardless of timing is worth pushing back on. So is a contract with no substitute speaker provision. Verbal assurances of "we'll work something out" don't protect you when the speaker goes dark two weeks out.
Content and recording rights: Who owns the recording? Can you use clips in post-event communications? Can the recording be included in member-only content? These questions become expensive if they are not answered before signing. AI speakers who are still employed at major technology companies often need to route presentations through their employer's legal or communications teams. The red flag is when neither the speaker nor their management raises this during initial conversations, and you discover it when you are asking for slides three weeks before the event.
How the bureau or agent makes money: Traditional speaker bureaus often add a commission to the speaker's fee without disclosing it. This means the quoted figure may not reflect what the speaker actually receives, and you have no clear way to know whether a speaker is being recommended because they are the right fit or because they generate the highest margin. Bureaus that won't explain their fee structure upfront deserve skepticism.
Crimson Speakers charges speakers a flat fee to be listed, which means working with us is always free for event planners. That structure removes commission incentives from speaker recommendations entirely. You can review how the model works if you're comparing bureau options.
References: What to Ask and What to Notice
References can tell you what you won't see in a reel or proposal: how the speaker behaved backstage, whether they were prepared, how the audience responded to the parts that weren't highlights.
Warning signs in the reference process start before the call:
- References are only provided after you've signed a letter of intent
- The list is concentrated in a single period (suggests recent bookings haven't gone as well)
- References appear to be managed relationships rather than independent clients
On the call, move past "was she good?" and ask questions that require specific recall:
- What was the audience's biggest question during Q&A, and how did the speaker handle it?
- Was there anything in the presentation that felt dated or that the audience pushed back on?
- Did the speaker's team communicate clearly before and during the event, or were there last-minute requests?
- Would you book them again for the same audience, or for a different one, and why?
A reference who can't answer these questions specifically, or who sounds like they are reading from the speaker's marketing materials, isn't giving you useful information.
Pre-Contract Red Flag Checklist
Use this before signing any AI speaker agreement:
Speaker vetting:
- Have you watched a full-length, unedited recording from a comparable event?
- Can the speaker name specific events, dates, and audience types, not just logo categories?
- Does the speaker's domain expertise match your audience's industry and knowledge level?
- Did the discovery call include substantive questions about your specific audience context?
- Can the speaker describe what falls outside their expertise?
Content and customization:
- Has the speaker requested audience-specific information beyond the conference theme?
- Is there a defined customization process with a timeline?
- Is the speaker current on AI tools and limitations, not referencing capabilities that are two years old as cutting edge?
Technical and logistics:
- Did the rider arrive before the contract?
- Does the speaker have a documented backup plan for technical failures?
- Have you confirmed rider requirements with your venue's AV team?
Contract and commercial terms:
- Are cancellation terms tiered and reasonable for both parties?
- Are content ownership and recording rights explicitly defined?
- Does the contract include a qualified substitute speaker provision?
- Do you understand how the bureau or agency makes money?
- Has the speaker confirmed whether employer review of content is required, and provided a timeline?
References:
- Are references from recent events (within 18 months) with comparable audiences?
- Have you asked questions that require specific recall, not just general endorsements?
Most Problems Are Predictable
Most AI speaker problems at events aren't surprises. They are the predictable result of a warning sign that was present before the contract and wasn't acted on. The speaker who couldn't handle substantive Q&A showed you that in the discovery call, when they sidestepped the one pointed question you asked. The rider conflict that ate two hours of setup time was foreshadowed by the team who said "we'll sort out the tech stuff closer to the date."
Protecting your event means developing the pattern recognition to catch those signals before they become problems. A speaker who earns your trust during the vetting process is far more likely to earn your audience's trust on stage.
If you'd like help evaluating AI speaker candidates, the team at Crimson Speakers works with planners at every stage of the booking process and can help you vet options against your specific audience and objectives.
Related planning resources
Use these Crimson Speakers planning resources to connect this decision to the next booking step:
- AI Strategy Speakers for audiences that need a practical business transformation keynote.
- How It Works for the intake, shortlist, and booking process.
- Request a Speaker when you are ready to compare available AI keynote options.