When Pricing Agents Meet Buying Agents: Personalized Pricing and Verifiable Trust

Same fairness rule, different posterior, no trade.

We study personalized pricing when seller and buyer principals delegate to agents that receive different value signals and execute machine-enforced mandates.

Equal nominal surplus rules can be incompatible because each is applied to its agent's posterior.

In one common environment, we solve nested pricing, inspection, mandate-selection, and repeated-relationship subgames.

Signal attestation and execution attestation have different effects: evidence of a high seller signal can legitimate a high price, whereas evidence of a restrained rule can prevent unauthorized extraction.

Verification can recover trade and soften both principals' policies only when it covers the disputed proposition.

In repeated transactions, attributable offers above a buyer's reference cap depreciate relationship capital, producing a state-dependent Markov policy and a reference-respecting region.

Verification is privately underprovided when the seller does not internalize avoided buyer inspection and relationship spillovers, but can be overprovided when it facilitates extraction.

The model links AI measurement, algorithmic extraction, and verifiable restraint without treating trust as software emotion or verified execution as proof of true value.