Anthropic created a test marketplace for agent-on-agent commerce

In a recent experiment, Anthropic created a secret marketplace where AI agents represented both buyers and sellers, making real deals for real goods and real money.

The company relented this test – which it called Project Deal – was just “a pilot experiment with a self-selected participant pool” of 69 Anthropic employees who were given a $100 budget (paid out via gift cards) to buy stuff from their coworkers.

Still, Anthropic said it was “impressed with how well Project Deal worked,” with 186 closed deals worth a total of more than $4,000.

The company said it was essentially running four separate marketplaces with different models: one that was “real” (where everyone was represented by the company’s most advanced model, and where deals were actually honored after the experiment) and another three for study.

Apparently, when users are represented by more advanced models, they “get objectively better results,” according to Anthropic. But users didn’t seem to notice the disparity, raising the possibility of “gaps in agent quality” where “people on the losing side may not realize they are worse off.”

Furthermore, the initial instructions given to the agents did not appear to affect the likelihood of a sale or the negotiated prices.