How to Find an AI Keynote Speaker Who Actually Fits Your Event
Finding the right AI keynote speaker comes down to one filter that removes most of the field: book someone who has shipped an AI system to production recently, not someone who only narrates the trend from a slide deck. A practitioner can field a hard question from your engineers, tailor the talk to your audience, and send the room home with something specific to act on. To find them, scan the speaker line-ups of practitioner events, ask engineering leaders you trust for names, and check each candidate's public work: recent talks, writing, and open-source activity. Shortlist three, run a short scoping call with each, and choose the one who pushes back on your brief with useful corrections rather than agreeing to everything.
I'm Mahmoud Zalt, an AI systems architect with 16 years building production software. Through Sista AI I help teams take AI from pilot to production, and I speak at conferences and team events on that same work.
Where to Look First
The best speakers rarely advertise on generic bureau sites. They show up where practitioners gather. Start with these channels, roughly in order of signal quality:
- Referrals from engineering leaders. Ask two or three technical people you respect who they have seen give a talk that actually taught them something. A warm referral beats any directory.
- Practitioner conference line-ups. Events like the AI Engineer Summit, QCon, and strong regional meetups accept speakers through technical review, not name recognition. Browse recent programs and watch the recorded sessions.
- Public work. Search candidates on GitHub, YouTube, and their own writing. Someone who reasons through a real problem in public is showing you exactly how they will handle your stage.
- Applied-AI communities. Slack and Discord groups focused on shipping AI often surface who gave a memorable internal talk.
Speaker bureaus are fine for logistics once you know who you want, but they optimize for bookability and brand, not for depth. Use them to close a deal, not to discover talent.
How to Vet a Shortlist in One Call
Once you have three names, a single 20-minute call tells you most of what you need. Ask each of these, and listen for specificity:
- What did you ship in the last year? You want a concrete system, a team size, and an outcome. A vague 'I advise organizations on AI strategy' is a warning sign.
- What went wrong and how did you catch it? Anyone who has run AI in production has a real failure story: retrieval drift, a bad model upgrade, runaway tool calls. The story should be specific.
- Can you handle live questions from senior engineers? A good speaker welcomes this and can describe what a hostile question looks like and how they would answer it.
- What would you tell us not to build right now? The ability to say no to a specific pattern, with a reason, is the clearest signal of real judgment.
Match the Speaker to Your Audience and Format
The right person depends on who is in the room. A talk that lands with executives will bore senior engineers, and the reverse is just as true. Use this as a rough guide:
| Audience | Best format | Optimize for |
|---|---|---|
| Executives and product leaders | Keynote plus Q&A | Strategy, risk, build vs buy, governance |
| Mixed conference crowd | Single 40-minute talk | Concrete examples and opinionated, quotable takeaways |
| Engineering team | Talk or hands-on workshop | Architecture patterns, evals, real production detail |
| Internal all-hands | Fireside or panel | Relatable examples from your own domain |
Decide the outcome you want before you decide the format. If you want registrations, a recognizable name helps sell tickets. If you want your team to think differently on Monday, depth beats fame every time.
Frequently Asked Questions
How do I find a good AI keynote speaker?
Start with referrals from engineering leaders you trust, then check practitioner conference line-ups and each candidate's public work. Shortlist three, run a short scoping call, and pick the one who improves your brief instead of just accepting it. Recent talks, writing, and open-source activity are stronger signals than a polished speaker page.
What should I ask a speaker before booking?
Ask what they shipped in the last year, what went wrong and how they caught it, whether they can handle live technical questions, and what they would advise you not to build right now. Specific answers point to a practitioner; general talking points point to a pundit.
How far in advance should I book?
For a single remote talk or podcast, two to three weeks is usually enough. For a workshop with content built around your team, plan four to six weeks. For an on-site keynote that involves travel, eight to twelve weeks is safer.
Should I book a famous name or a practitioner?
For a technical audience, almost always the practitioner. A famous name helps sell tickets to a broad event, but a practitioner who has run real systems will give your engineers takeaways they still use months later.
Book a Speaker Who Can Answer the Hard Questions
The right AI keynote speaker is the one your audience still quotes months later, because they said something specific and true that changed how the room thinks about a real problem. That person is usually not the biggest name on the roster; they are the one who has actually shipped the work they are talking about.
My Public Speaking service covers talks, workshops, and podcasts on AI systems, architecture, and engineering leadership. Remote talks start at $1.8K, a half or full-day workshop is $3.9K, and on-site keynotes run $4.8K to $9K plus travel. If you are planning an event and want someone who can field your engineers' toughest questions, see the details and reach out.
