Board Case Study · Consumer (Retail & Travel)

The Price That Set Itself

The pricing AI learned to maximize margin, and to mirror competitors within minutes. A regulator calls it tacit algorithmic collusion. No human ever agreed to fix prices.

The situation

A public consumer company uses a reinforcement-learning pricing engine across millions of SKUs. It lifted margin and was a board-level success story. A regulator has now opened an inquiry: the model appears to have learned to match competitors' price moves within minutes, producing supra-competitive prices that look like coordination.

No executive ever instructed the system to collude, and competitors use similar third-party engines trained on overlapping signals. Antitrust law is adapting to algorithmic pricing, and "the model did it on its own" is an unsettled defense. Consumer-harm and disclosure questions follow close behind.

The board must decide what it knew or should have known, whether to constrain the model's behavior, and how to explain an outcome no person intended but the company benefited from.

“Nobody told it to collude. It learned that mirroring our competitors pays. Is that the model's behavior, or ours?”

Chair, Risk Committee

The decision on the table

  • Assess antitrust exposure when an autonomous pricing model produces coordination no human directed.
  • Decide what constraints, monitoring, and explainability the model now requires.
  • Set the board's duty to understand and oversee algorithmic systems that act in regulated markets.

Run this case with your board.

The full case, scenario architecture, board materials, and facilitation notes, is shared with boards and partners on request.

Request the full case All case studies