Deciding in the Age of AI, Lessons from a Poker Player
Keeping the edge when AI reshuffles the deck.
A four-step decision model, and why copying the AI is a strategic mistake.
A keynote about decision-making, risk management, and the place of humans alongside artificial intelligence.
Poker is the only laboratory where the quality of a decision can be measured independently of its outcome. That is what makes it a training ground rather than a casino anecdote.
When the world's 100 best poker players faced the machine, and lost, they learned a lesson that applies to every company in transformation: AI gives you the state of the art, but it can neither understand you nor decide for you.
Participants leave with a concrete four-step decision model, Learn, Reason, Act, Relearn, applicable the very next day.
Depending on your needs, the talk goes hands-on: AI agents serving business teams, vibecoding to prototype without waiting for engineering, and framing of real use cases. The European AI Act has required every organisation since February 2025 to ensure its people reach a sufficient level of AI literacy, an obligation these formats help meet.
"96% of poker players lose over time. Chance has almost nothing to do with it. The 4% who win simply have a method for deciding."
Key takeaways
- 1
The quality of a decision ≠ its immediate outcome (tool: Sklansky Dollars)
- 2
Two kinds of errors, including the invisible "leaks" that cost the most
- 3
Copying the AI is a strategic mistake (Game Theory Optimal)
- 4
Three roles for the augmented employee: operate the tool, assess its answer, own the result
What the room takes away
The four-step model
Four moves replayed at every hard decision. This is what participants write down, and what they reopen the next morning.
- 1 Learn Gather the state of the art, AI included.
- 2 Reason Weigh the options, look for what is missing.
- 3 Act Choose, then set things in motion.
- 4 Relearn Judge the decision, not only its outcome.
and round again
Two ways to work together
From inspiration to transformation. Place the cursor where you stand.
A keynote that lands
The stage format: storytelling, live experiments with the room, demonstrations. Your teams leave with a decision model and the urge to use it.
Book the keynote →A programme that sticks
The operational follow-through: AI culture, tooling choices and rollout, process redesign, upskilling your teams. Here we stop telling and start installing.
Discuss the programme →Available as
Audiences : Executive committees · conventions · digital transformation events · partner days
They moved forward on AI with Nicolas
This keynote at your event?
The two other keynotes : Reading the Table · Capturing Attention