You booked the AI session last year. The speaker was polished, the slides were impressive, and the survey scores came back fine. Then you stood in the hallway afterward and listened to an agency principal ask the question that stuck with you: so what do we actually do?
Six months later, nothing about how your members work had changed. The producers still spend their evenings re-keying quotes. The service teams still drown in certificate requests. And now you are the one choosing the AI speaker again.
The bar for this booking is simple to state and hard to meet: will anything be different on the Monday after the conference? Everything below follows from that question. These are the criteria I would tell you to use even if you never book me - and before we are done, I will tell you plainly when you should not.
What separates a working session from a good show
They have worked inside insurance, not just spoken near it. Generic AI material translated at the door falls apart the moment the first hand goes up, because in your rooms, the first hand asks about compliance. A speaker who belongs on your stage can talk about carrier systems, the AMS, licensing, E&O exposure, and what client health information can and cannot touch. The test is simple: ask them to describe a producer's week. When quoting, re-keying, certificates, renewals, and claims paperwork come up without your prompting, you are talking to someone who has been in the room.
They speak from delivered work. A slide deck about the future is cheap to build. A session built on rooms the speaker has actually served is not. Ask about the last insurance audience they stood in front of and what happened after. A practitioner answers with specifics - who was in the room, what got built, what the follow-up looked like. A futurist answers with trends.
The room leaves with something to do. Inspiration decays by the drive home. Skills do not. Ask what your members will have open on their laptops during the session, and what will exist at the end of it that did not exist at the start. If the honest answer is a feeling, keep looking.
They can hold a mixed room. Your audience is agency principals, producers, and service staff sitting in the same rows, and their daily friction differs sharply. A principal is thinking about governance and E&O. A producer wants hours back in front of clients. Service staff want the certificate requests to stop breeding. Ask the speaker how the session lands differently for each seat. If the answer amounts to the message being universal, the friction was never considered.
The questions to ask on the call
Selection happens on the discovery call, and the right questions make the wrong speaker visibly uncomfortable. Use these.
Walk me through the last insurance room you spoke to. What did they build? You are listening for names, dates, and outcomes offered readily. A vague nod toward financial services clients means the insurance experience is thinner than the website suggested.
What would you tell our agents about putting client health information into a chatbot? Any speaker can wave at a disclaimer. The right one will talk about governance - which tools hold what data, what an agency AI policy should actually say, and why ungoverned Shadow AI use is an E&O problem waiting for a claim.
What will our members have on their screens by the end? The answer separates a talk from a working session. Both have a place. You should know which one you are buying before you sign.
How do you handle the producer who tried ChatGPT and decided it was not worth the time? That producer is in every insurance room, and they are usually right about the experience they had. A good speaker treats that scar as the starting point of the session, because the problem was never the producer. The problem was bolting a tool onto a workflow nobody redesigned.
What do you need from us before you will take the stage? This one is quietly revealing. A practitioner asks about your rooms: who is attending, what their week looks like, which workflows hurt the most. A show speaker asks about the AV setup.
What should our members be doing differently thirty days later, and how would we know? When a speaker has a real answer here, follow-through was designed into the session rather than hoped for afterward.
Take this into the call
Print one per candidate. Note what they actually said, not how it felt. Compare the sheets, not the impressions.
- The last insurance room they spoke to, and what it built.Listening for: names, dates, outcomes offered readily. Vague sector references mean thinner experience than the website suggested.
- What they would tell your agents about client health information in a chatbot.Listening for: governance specifics and E&O exposure, not a disclaimer.
- What members will have on their screens by the end.Listening for: an artifact, or an honest admission that this is a talk.
- How they handle the producer who tried AI and quit.Listening for: the workflow as the failure, never the producer.
- What they need from you before taking the stage.Listening for: questions about your rooms. AV-only questions are the tell.
- What should be different thirty days later, and how you would know.Listening for: follow-through designed in, not hoped for.
Candidate: ______________________ · Date: __________ · Go / No-Go
When you need a different kind of speaker
Sometimes the answer is genuinely someone else, and pretending otherwise would make everything above suspect.
If the session's job is energy - a high-momentum opener before the expo hall, a celebration keynote at the awards dinner - book an inspirational speaker and let them do that craft well. It is a real craft. It just is not mine, and a hands-on builder in that slot will underdeliver on the lift you actually need.
If your agencies are already past adoption and need deep technical work - carrier API integrations, AMS engineering, custom deployments - you need a technical consultant on a project engagement, not a stage session. A conference slot is the wrong container for that work no matter who fills it.
And if the budget conversation keeps coming back to how the room will feel rather than what the room will do afterward, be honest with yourself about which purchase you are making. Neither is wrong. They are different purchases.
Match the format to the job you are hiring it to do
| Inspirational keynote | Practitioner working session | Technical consultant | |
|---|---|---|---|
| Book it when | You need lift: an opener before the expo hall, an awards dinner, a morning after a hard year. | Your members have heard about AI and still work the way they did last year. | Adoption already happened and the constraint is systems: carrier APIs, AMS, integrations. |
| Slot it fits | Main stage, 45 to 60 minutes, no laptops. | Keynote plus a hands-on block, or a half-day breakout. | Not a conference slot at all. A project engagement. |
| What the room gets | Energy, permission, a shared reference point for the hallway. | A first assistant built during the session, on their own work. | Architecture, scope, and a build plan for the agency that hired it. |
| What exists Monday | A memory, and usually a few forwarded slides. | Something running that takes a slice of paperwork off a desk. | A statement of work, not a room-wide behavior change. |
| How to brief it | Theme, tone, and the mood you want in the room afterward. | Who is attending, what their week looks like, which workflows hurt most. | Systems inventory, data boundaries, and who owns the outcome internally. |
| Where it fails | Nothing changes, and next year you book the same slot again. | Wrong room: an audience that wanted a show sits through a build. | Wrong container: deep implementation work compressed into a stage hour. |
Status: all three are legitimate purchases · know which one you are signing
What a good outcome looks like
Judge the session by the weeks after it, because that is when the applause has faded and enrollment season has not.
The insurance version of this problem has a name: Admin Drag - the quoting, the re-keying into the AMS, the certificates, the renewals, the claims paperwork that swallows most of a producer's week. A working session gets a room to see its own version of that drag clearly, then starts members building the first piece of a Digital Crew - an assistant that takes a real slice of the paperwork off a real person's desk.
That outcome is observable. This August at Berwick Insurance Group, Desiree Bayne, their Sales Retention Manager, described agents leaving the workshop with working AI assistants they could use that same day. And after a session with insurance agents at the Kellogg Business Summit in May, Mike Welker of Insurance Center of Utah wrote to me two months later to say the presentation had changed his life. Sixty days is long past the applause. That is the window that tells you whether a session worked.
When you are checking a speaker's references, ask them the same question: what changed in the sixty days after?
Questions planners ask
How do I know if an AI speaker really knows insurance?
Ask them to describe a producer's week. Speakers who have worked inside the industry bring up quoting, re-keying into the AMS, certificates, renewals, and claims paperwork without prompting, and they can speak to compliance, carrier systems, and E&O exposure. Then ask about the last insurance room they served and what that audience built afterward.
What is the difference between an AI keynote and a working session?
A keynote delivers ideas and energy. A working session leaves attendees with something built. Ask what members will have on their screens by the end. Both formats have a place, but you should know which one you are buying before you sign.
Should we worry about compliance if our agents start using AI after the session?
You should ask about it before the session, not after. A credible speaker addresses governance directly: which tools hold what data, what an agency AI policy should cover, and why ungoverned Shadow AI use creates E&O exposure. A session that sends agents off experimenting without that framing has handed you a risk rather than a capability.
How should a speaker handle agents who already tried AI and gave up?
That producer is in every insurance room, and their experience was real. A good speaker starts from that scar rather than dismissing it, showing why a tool bolted onto an unchanged workflow failed and what redesigning the work first changes.
What should we expect in the thirty to sixty days after an AI session?
Members acting on what they built: assistants in daily use, paperwork moving off desks, follow-up questions coming in. Ask the speaker how follow-through is designed into the session, and ask their references what changed in the sixty days after theirs.
When should we book a different kind of AI speaker?
Book an inspirational keynoter when the job is energy, such as an opener before the expo hall. Engage a technical consultant when agencies need deep integration work like carrier APIs or AMS engineering. A practitioner-led working session serves rooms that need agents and staff working differently the following week.
Checking dates
If you are planning an insurance event and want to see whether the calendar lines up, everything is on the speaking page, including what I cover with insurance and Medicare distribution audiences. I will be on stage at WAPCON this September - if you will be there, come see a session in person before you decide anything.
