FutureUP workshop at the PPS profitABLE26 Conference, 2 Dec 2026
Most AI pricing initiatives fail commercially, not technically
When an AI pricing project fails, the "usual suspect" is a broken algorithm. But in most cases, the problem is commercial, not technical. A wrong or unclear objective. Data that looked valuable and ended up adding noise. Or a recommendation that was probably right and never got executed, because it wasn't aligned with operational processes or nobody trusted it.
That gap — between a model that works and a model that changes what a business actually does — is what this workshop is about.
Four lessons the real-world evidence keeps confirming
Frame the decision before choosing a model.
"Which AI solution should we use?" is the wrong first question. The right one is "which decision are we trying to improve, and under what constraints?" — because that answer settles everything downstream: what data you need, which method fits, what the output has to look like, and how you will know it worked.
Choose wisely here — otherwise, AI will solve the wrong problem perfectly.
More data is not automatically better.
On one engagement, competitor prices seemed like an obvious addition to the working predictive model. The data was clean: only a handful of relevant players, comparable offerings, and known prices. Still, accuracy deteriorated significantly once they were added. The problem wasn’t the data. Competitor prices simply didn’t contribute enough new signal beyond what the model had already captured.
Commercial sense should come first; then allow predictive accuracy to decide what earns a place.
Validation is what earns the right to act.
Before anyone bets margin on a recommendation, they should see the model predict unseen outcomes and customer behavior. Not an uplift estimate, but evidence readable by business people, not data scientists.
That changes the conversation from "trust the AI and its price suggestions" to "the model successfully predicted real sales performance under changing conditions — here is what that implies for future prices and outcomes."
The best analytical recommendation is not the best operational one.
A recommendation nobody acts on is worth nothing. Salespeople fear customer pushback. Managers worry about losing market traction. Whoever sponsors the decision fears the risk.
AI does not remove the risk. It makes it an informed one by providing the reasoning behind pricing decisions, expected business impact and confidence, and sales scenarios with trade-offs. But its recommendations still need a realistic path to implementation to change anything.
The workshop
Where AI Price Optimization Fails and Succeeds: Real-World Lessons
Workshop 6 · Wednesday 2 December 2026 · 9:00–16:00 · Mövenpick Hotel Amsterdam City Centre
A full-day, hands-on session for pricing leaders — built for the people who commission, evaluate, and govern AI pricing. Real-world lessons from B2B and key differences from B2C.
Every session pairs a lesson with practical work. You will leave knowing:
How to frame a pricing decision — objective, constraints, scope — before any model is chosen.
How to match the level of AI capability to your business needs, and incorporate it smoothly into everyday operations.
How to evaluate whether an AI pricing model can be trusted, using specific validation evidence.
It draws on real-world evidence from AI-driven price optimization engagements, including validations that held up through the difficult 2026 market.
The workshop also carries CPP credit toward the Certified Pricing Professional designation — the globally recognized certification of pricing excellence, and a network now nearly 3,000 CPPs strong.
This year's conference additionally marks the launch of the PPS Certified Pricing & Profit Leader (CPPL) designation and the PPS Leadership Academy.
Register for profitABLE26 Amsterdam ↓
AI-driven pricing insights to unlock hidden revenue & margin
Using your data. Minimal involvement. No implementation needed.
Comments