A speaker guide for event organizers and planners
Choose a session that moves the work forward.
Choose an AI adoption session by the decision your audience needs to make. Then brief the speaker with a real workflow, a fair baseline and a realistic next step.
Start with the decision, not the name.
A keynote can create shared language. A working session can help define a bounded experiment. Neither automatically produces an implementation plan or proves a financial return. Confirm what a speaker actually offers rather than inferring it from a biography.
| If the audience needs to | Discuss this format | Ask for a realistic output |
|---|---|---|
| Understand what AI changes about work | Keynote with relevant examples and Q&A | A shared vocabulary and questions to investigate, not a deployment guarantee |
| Choose one workflow to test | Facilitated adoption workshop | A bounded task, baseline, quality criteria and accountable owner |
| Challenge a proposed investment | Executive discussion or decision clinic | Visible assumptions, missing evidence and a stop or continue decision process |
Format availability, fees and engagement terms must be confirmed directly. A university or organization profile is evidence of expertise, not a booking offer.
Five questions that make a shortlist useful.
- Can the speaker connect to this audience’s work?Share the audience roles, technical familiarity and one representative workflow. Ask for a current outline and a talk sample relevant to those people, not just a general highlight reel.
- What evidence sits behind the examples?Ask what was measured, over what period, with what comparison and what limitations. Distinguish a research finding, personal case study, illustrative scenario and forecast. Do not generalize one successful task to the whole organization.
- Does the proposed approach count the full workload?Probe preparation, source checking, review, correction and exceptions. A large reduction in drafting time may become a small reduction in approved-work time. Use the measurement guide to frame a fair comparison.
- Where does the speaker set boundaries?Look for an explanation of quality, authorized data use, human accountability and when not to automate. No average time saving compensates for an unacceptable data exposure or unauthorized action.
- What will participants leave ready to do?Agree on an output that fits the format: a question set, test plan or decision brief. Confirm who will follow through. Event satisfaction and attendance do not establish workflow improvement.
Guaranteed “10x” results, results without a baseline, demos presented as production evidence, unapproved use of attendee data, or an unclear distinction between speaking and implementation work. Ask for clarification; do not turn these signals into an invented speaker ranking.
Make the work before and after visible.
Before the session
Bring a bounded task.
Describe inputs, the approved output, current process and reviewer. Bring public or synthetic examples unless data use is specifically authorized.
Confirm duration, Q&A, delivery mode and time zone, travel, AV, rehearsal, accessibility, recording permissions and follow-up material.
During the session
Separate facts from assumptions.
Capture the proposed workflow, expected quality, missing evidence and stop conditions. Ask where examples do not apply to your context.
Record a next decision and an owner. Do not treat applause, confidence or a polished demonstration as outcome evidence.
After the session
Check the actual work.
Run a bounded, authorized test using comparable tasks. Record total human time, accepted or rejected outputs and exceptions.
Use the quality guide and capacity calculator. Decide whether to continue, revise or stop; report uncertainty.
Build a brief you can actually send.
Complete the fields, generate a plain-text brief and review it before sharing through the speaker’s official inquiry route. This tool does not contact anyone, book a speaker or submit your inputs.
Use the directory as a starting point.
The speaker directory contains ten sourced profiles, alphabetically ordered by first name. Compare documented topics and ask the briefing question on each card. Inclusion is not a ranking, endorsement, availability check or independent test of delivery quality.
The questions and brief builder on this page are editorial planning tools. Risk and evaluation concepts draw on the NIST AI Risk Management Framework and Microsoft evaluation guidance; the tools are not endorsed by those organizations or by listed speakers. Read the method and limits.