2
3
5
AI agents are increasingly used in consumer-facing applications to assist with tasks such as product search, negotiation, and transaction execution. In this paper, we investigate a future setting where both consumers and merchants authorize AI agents to automate the negotiations and transactions in consumer settings. We aim to address two questions: (1) Do different LLM agents exhibit varying performances when making deals on behalf of their users? (2) What are the potential risks when we use AI agents to fully automate negotiations and deal-making in consumer settings? We designed an experimental framework to evaluate AI agents' capabilities and performance in real-world negotiation and transaction scenarios, and experimented with a range of open-source and closed-source LLMs. Our analysis reveals that deal-making with LLM agents in consumer settings is an inherently imbalanced game: different AI agents have large disparities in obtaining the best deals for their users. Furthermore, we found that LLMs' behavioral anomaly might lead to financial loss when deployed in real-world decision-making scenarios, such as overspending or making unreasonable deals. Our findings highlight that while automation can enhance transactional efficiency, it also poses nontrivial risks to consumer markets. Users should be careful when delegating business decisions to LLM agents.
Our framework simulates real-world negotiations between two AI agents - one buyer and one seller - over consumer products. While both know the retail price, only the seller knows the wholesale cost. The buyer operates within a budget limit, and negotiations follow strict rules: sellers can't go below wholesale cost, and buyers can't exceed their budget. Through multiple rounds of offers, sellers aim to maintain retail prices while buyers seek the best discounts.
From Model Capability Gap to Economic Loss. Our analysis reveals that when agents with different negotiation capacities compete against each other, users who employ weaker agents experience significant economic losses.
Economic impact of model imbalance in agent negotiations. We analyze seven model pairings with successful negotiation overlaps. Using DeepSeek-R1 vs. DeepSeek-R1 as baseline. From this table, we can see that whether a strong buyer faces a weak seller or vice versa, the party with the weaker agent suffers a strategic disadvantage resulting in economic loss.
From Model Anomaly to Economic Loss. We identified four types of negotiation anomalies in our experiments: (1) Constraint violations where agents ignore budget limits or wholesale costs; (2) Excessive payments where buyers offer above retail prices; (3) Negotiation deadlocks where agents get stuck in endless loops; and (4) Early settlements where high-budget buyers compromise too quickly instead of negotiating better deals.
@article{your_paper_citation,
title={The Automated but Risky Game: Modeling Agent-to-Agent Negotiations and Transactions in Consumer Markets},
author={Shenzhe Zhu and Jiao Sun and Yi Nian and Tobin South and Alex Pentland and Jiaxin Pei},
journal={},
year={2025}
}