The call comes in three days before the conference: the AI keynote speaker's laptop runs exclusively on Thunderbolt 4, your venue AV company only has HDMI adapters at the stage, and the "live AI demo" segment requires a hardwired internet connection the ballroom was never wired for. The speaker's green room is on the second floor. The main stage ethernet drop is on the first. No one mentioned any of this during contracting.
This is not a hypothetical. It is the kind of scenario veteran event directors walk into with AI and tech-focused speakers, and it is entirely preventable if you know what to ask for before you sign the contract.
AI keynote speakers have more complex AV requirements than virtually any other speaker category. A business author touring a book needs a lectern, a clicker, and 1080p display output. An AI speaker doing a live demonstration of a language model, a robotics platform, or a real-time data visualization may need a dedicated ethernet line, three separate display outputs at specific resolutions, audio routing from their laptop through the house PA, and a backstage setup that mirrors the stage environment exactly. The gap between what a standard AV rider covers and what an AI speaker actually needs is where conferences get into trouble.
Why AI Speakers Have Different Technical Needs
The defining characteristic of an AI keynote is the live demonstration. Unlike a traditional keynote where slides are the centerpiece, AI speakers frequently run actual software during their presentation: language models generating text in real time, computer vision systems processing live video feeds, custom-built tools that interact with live APIs. Each of these creates a dependency chain a standard AV setup was not designed to support.
The other distinguishing factor is resolution sensitivity. AI demos often involve text-heavy interfaces, data tables, code output, or fine-grained visual outputs. A blurry or low-contrast display does not just look bad; it makes the actual content of the demo illegible to the audience. This is fundamentally different from a slide deck, where a slightly degraded image is a cosmetic problem.
These two factors, live software dependency and high-fidelity display requirements, frame everything else in this guide.
The Display Checklist Every Event Planner Needs
Before confirming an AI speaker's booking, walk through this checklist with your AV company:
Output ports and adapters:
- Confirm what video outputs the speaker's laptop has (USB-C/Thunderbolt, HDMI, Mini DisplayPort)
- Confirm what your venue's signal chain accepts at the first input point
- Request that the speaker provides their own adapter, and have the AV company keep a matching backup on-site
- Confirm the adapter supports the required resolution and refresh rate, not just a physical connection
Resolution and aspect ratio:
- Request the speaker's slide deck aspect ratio in advance (16:9 is standard; some AI demo environments run native 16:10 or require specific canvas dimensions)
- Confirm your projection or LED wall can display at the speaker's native resolution without upscaling that muddies text
- For large LED walls, ask your AV company specifically about pixel pitch and minimum readable text size at the content resolution
Confidence monitors and stage monitors:
- AI speakers doing live demos frequently cannot look up at the main screen while operating their laptop
- Confirm there is a stage or confidence monitor positioned where the speaker can see slide notes and a mirrored output of their laptop screen simultaneously
- Some speakers will request a separate monitor for their laptop display, so the audience screen shows only what they choose to push there
Signal chain testing:
- The signal chain test needs to happen with the actual laptop that will be used during the presentation, not a substitute
- This is non-negotiable for live AI demos. Codec differences between machines affect video playback, and some AI interfaces behave differently under display scaling settings
Internet Connectivity: The Non-Negotiable
Conference wifi is not a viable connection for a live AI demo. The reasons are straightforward: shared bandwidth with hundreds of concurrent users, variable latency that causes API calls to time out mid-demonstration, and no QoS guarantees even on "speaker priority" networks.
The correct setup for an AI speaker doing live demonstrations requires either a hardwired ethernet connection at the stage or a dedicated LTE/5G hotspot with a strong signal, treated as the backup rather than the primary.
Wired ethernet at the stage is the ideal. In older hotels and convention centers, the ethernet drops may only exist at the perimeter of a ballroom, not at stage center. If your speaker is doing a live demo, confirm this during your venue walkthrough, not after load-in. Running a cable from the wall to stage center requires conduit or a floor cable cover path, and that costs time and labor.
When wired ethernet is not available, the best practice is a dedicated cellular hotspot that belongs solely to the speaker's device during the presentation window. The speaker should bring their own device with data loaded; do not rely on the venue to provide this. Carriers vary significantly by venue location, so if the conference is in a concrete-walled convention center basement, the speaker may need to test signal in that specific room.
One practical detail that is easy to miss: if the AI demo requires back-end authentication or specific network whitelist rules, those need to be tested in the actual venue network environment. A speaker who has demoed flawlessly from a home office may run into corporate firewall rules the moment they plug into a hotel network.
Audio Routing for Interactive Demonstrations
Most AI speaker AV riders address microphone requirements but overlook audio output from the laptop. If the speaker's demo produces audio, such as a voice synthesis demo, an AI music generation tool, or a video clip that is part of the demonstration, that audio needs to travel through the house PA system, not just play from the laptop speaker.
The correct path is a stereo DI (direct input) box connected from the laptop's audio output to the venue's mixing console. This is standard practice, but it needs to be specified in advance. The AV company needs to know to bring the DI boxes, the audio engineer needs to know there is a laptop channel to account for in the mix, and a level check needs to happen with the actual demo audio, not a test tone.
Microphone selection for AI speakers is worth discussing directly with the speaker. Speakers who move around a stage while interacting with a laptop tend to prefer a headset or lavalier rather than a handheld, since their hands need to be free. Even so, some speakers have a strong preference based on what they find comfortable to wear over a 45-minute keynote. Get this information during the contracting phase, not the morning of the event.
Backstage Setup and Prep Room Requirements
The requirement that gets missed most often in AI speaker coordination is the backstage or green room setup. A speaker doing a live AI demonstration needs to run through the full demo sequence, including the parts that interact with live APIs, from backstage before going on stage. This means the prep room needs:
- An internet connection with the same bandwidth characteristics as what will be available on stage
- A display output that matches the stage setup (resolution, cable type)
- Enough time built into the run-of-show for a proper tech check, not a five-minute scan of the slide deck
The backstage tech check for an AI speaker should be scheduled for at least 30 to 45 minutes before the session, separate from any on-stage soundcheck. This is the session where the speaker actually runs the demo in conditions as close to stage conditions as possible. If an API token is going to expire or a network dependency is going to behave unexpectedly, you want to discover it backstage, not during the live presentation.
Bureaus like Crimson Speakers typically flag this requirement during the booking process so it gets written into the event timeline early. Backstage coordination is much easier to negotiate before the run-of-show is finalized than after.
What the Speaker Rider Should Specify
Speaker riders for AI and technology keynotes should be more detailed than a standard speaker rider. If you receive a rider that covers only display output and microphone type, ask the speaker's representation for a supplemental technical requirements document.
A complete technical rider for an AI keynote speaker should cover:
- Laptop model and operating system (so the AV company knows what adapters to have in inventory)
- Display output type and required resolution
- Whether the speaker will push one or two signals to the AV system (one for the confidence monitor, one for the house)
- Internet connectivity requirement and backup plan
- Audio output requirements (stereo DI, specific sample rate if relevant)
- Whether any external devices will be connected (webcam, external GPU, peripheral hardware for a robotics or AR demo)
- Backstage tech check timing requirement
- Load-in time requirement (some speakers need to arrive hours before the session to validate the full technical chain)
The rider is where you negotiate, not the morning of the event. If a rider item conflicts with your venue's capabilities, address it during contracting. It is far easier to say "we cannot provide hardwired ethernet at stage center; here is our proposed alternative" before the speaker is confirmed than to discover the conflict the day before the conference.
Preparing Your AV Company
Your AV company has set up hundreds of standard keynote presentations. Fewer of them have done a live AI demo at scale. The briefing you give them matters.
Walk them through the specific demo sequence if the speaker will share it. Explain which segments involve live internet connectivity and which can run offline if needed. Ask the AV director specifically about their experience with laptop-based live demonstrations, not just keynote presentations generally.
If your AV company has not done this type of setup before, consider requesting that the speaker's technical contact, which many AI keynote speakers bring, be part of the load-in and tech check process. That person knows exactly what the setup needs to look like and can troubleshoot in ways a venue AV technician unfamiliar with the specific software cannot.
One more thing worth clarifying with your AV company: who owns the problem if the technical setup fails mid-presentation. This is a contract question as much as a technical one. A clear escalation chain (speaker's tech contact, AV director, your event production contact) means that if something goes wrong on stage, everyone knows who to call and what they are authorized to decide.
Backup Planning Is Not Optional
Every live AI demo should have a recorded backup. This means the speaker has pre-recorded a clean walkthrough of the exact demo they plan to run live, exported at full presentation resolution, ready to play through the venue AV system if the live version fails.
This is standard practice for high-stakes presentations. It does not mean the speaker expects to fail; it means the speaker is a professional who has been on enough stages to know that dependencies fail unpredictably. Having the backup does not change the live presentation. It gives the speaker and your event team a graceful recovery path if something goes wrong.
The backup video should be loaded on a second laptop or on a dedicated media player in the AV rack, not on the same machine running the live demo.
For specific questions about technical riders and AV coordination during the booking process, the team at Crimson Speakers can connect you with speakers whose requirements are matched to your venue's capabilities, before a contract is signed.
Before your next event: Pull out your AV intake questionnaire and check whether it asks for video output type, internet requirements, audio output routing, and backstage tech check timing. If it does not, add those questions now, and save yourself the three-days-before phone call.