The Automated but Risky Game: Modeling and Benchmarking Agent-to-Agent Negotiations and Transactions in Consumer Markets
Open accessAbstract
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 explore a future scenario where both consumers and merchants authorize AI agents to fully automate negotiations and transactions. We aim to answer two key questions: (1) Do different LLM agents vary in their ability to secure favorable deals for users? (2) What risks arise from fully automating deal-making with AI agents in consumer markets? To address these questions, we develop an experimental framework that evaluates the performance of various LLM agents in real-world negotiation and transaction settings. Our findings reveal that AI-mediated deal-making is an inherently imbalanced game -- different agents achieve significantly different outcomes for their users. Moreover, behavioral anomalies in LLMs can result in financial losses for both consumers and merchants, such as overspending or accepting unreasonable deals. These results underscore that while automation can improve efficiency, it also introduces substantial risks. Users should exercise caution when delegating business decisions to AI agents.
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
Agents already perform, albeit imperfectly, tasks such as shopping, negotiation, or search for products on e-commerce sites.
“This possibility is near, as agents already perform, albeit imperfectly, tasks such as shopping (Dammu et al. 2025), negotiation (Zhu et al. 2025), or search for products on e-commerce sites (Zeff 2025;Allouah et al. 2025).”
From The Coasean Singularity? Demand, Supply, and Market Design with AI Agents (2025) · cites this paper as Zhu et al. 2025
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Cites
Scaling Laws for Neural Language Models
Kaplan, J., McCandlish, S., Henighan, T., Brown, T. B., Chess, B., Child, R., Gray, S., Radford, A., Wu, J., & Amodei, D. (2020) · arXiv · Cited by 1,749 · cited 1× · see citing works
“The scaling law of LLM suggests that model capabilities generally improve with increasing parameter count (Kaplan et al., 2020; Hoffmann et al., 2022; Bi et al., 2024; Zhang et al., 2024).”
The Landscape of Emerging AI Agent Architectures for Reasoning, Planning, and Tool Calling: A Survey
Masterman, T., Besen, S., Sawtell, M., & Chao, A. (2024) · arXiv · doi:10.48550/arxiv.2404.11584 · Cited by 35 · cited 1× · see citing works
“Recently, large language model (LLM) powered AI agents have demonstrated remarkable capabilities and are increasingly adopted for real-world tasks (Xu et al., 2024; Masterman et al., 2024).”
Flagged first, then by citations
Research works printed but not linked: 43
- Sahar Abdelnabi, Amr Gomaa, Sarath Sivaprasad, Lea Schönherr, and Mario Fritz. 2024. Cooperation, competition, and maliciousness: Llm-stakeholders interactive negotiation. Advances in Neural Information Processing Systems , 37:83548-83599.
- Henrik Agndal, Lars-Johan Åge, and Jens EklinderFrick. 2017. Two decades of business negotiation research: an overview and suggestions for future studies. Journal of Business & Industrial Marketing , 32(4):487-504.
- Jasper Bakker, Aron Hammond, Daan Bloembergen, and Tim Baarslag. 2019. Rlboa: A modular reinforcement learning framework for autonomous negotiating agents. In AAMAS , pages 260-268.
- Xiao Bi, Deli Chen, Guanting Chen, Shanhuang Chen, Damai Dai, Chengqi Deng, Honghui Ding, Kai Dong, Qiushi Du, Zhe Fu, and 1 others. 2024. Deepseek llm: Scaling open-source language models with longtermism. arXiv preprint arXiv:2401.02954 . · arXiv
- Federico Bianchi, Patrick John Chia, Mert Yuksekgonul, Jacopo Tagliabue, Dan Jurafsky, and James Zou. 2024. How well can llms negotiate? negotiationarena platform and analysis. arXiv preprint arXiv:2402.05863 . · arXiv
- Mert Cemri, Melissa Z Pan, Shuyi Yang, Lakshya A Agrawal, Bhavya Chopra, Rishabh Tiwari, Kurt Keutzer, Aditya Parameswaran, Dan Klein, Kannan Ramchandran, and 1 others. 2025. Why do multi-agent llm systems fail? arXiv preprint arXiv:2503.13657 . · arXiv
- Sanxing Chen, Sam Wiseman, and Bhuwan Dhingra. 2024. Chatshop: Interactive information seeking with language agents. arXiv preprint arXiv:2404.09911 . · arXiv
- Zichen Chen, Jiaao Chen, Jianda Chen, and Misha Sra. 2025. Position: Standard benchmarks fail-llm agents present overlooked risks for financial applications. arXiv preprint arXiv:2502.15865 . · arXiv
- Terrence Chong, Ting Yu, Debbie Isobel Keeling, and Ko de Ruyter. 2021. Ai-chatbots on the services frontline addressing the challenges and opportunities of agency. Journal of Retailing and Consumer Services , 63:102735.
- Zehang Deng, Yongjian Guo, Changzhou Han, Wanlun Ma, Junwu Xiong, Sheng Wen, and Yang Xiang. 2025. Ai agents under threat: A survey of key security challenges and future pathways. ACM Computing Surveys , 57(7):1-36.
- Silvia Feliu. 2001. Intelligent agents and consumer protection. International Journal of Law and Information Technology , 9(3):235-248.
- Daya Guo, Dejian Yang, Haowei Zhang, Junxiao Song, Ruoyu Zhang, Runxin Xu, Qihao Zhu, Shirong Ma, Peiyi Wang, Xiao Bi, and 1 others. 2025. Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning. arXiv preprint arXiv:2501.12948 . · arXiv
- Lewis Hammond, Alan Chan, Jesse Clifton, Jason Hoelscher-Obermaier, Akbir Khan, Euan McLean, Chandler Smith, Wolfram Barfuss, Jakob Foerster, Tomáš Gavenˇ ciak, and 1 others. 2025. Multiagent risks from advanced ai. arXiv preprint arXiv:2502.14143 . · arXiv
- John C Harsanyi. 1995. Games with incomplete information. American Economic Review , 85(3):291-303.
- He He, Derek Chen, Anusha Balakrishnan, and Percy Liang. 2018. Decoupling strategy and generation in negotiation dialogues. arXiv preprint arXiv:1808.09637 . · arXiv
- Dan Hendrycks, Collin Burns, Steven Basart, Andy Zou, Mantas Mazeika, Dawn Song, and Jacob Steinhardt. 2020. Measuring massive multitask language understanding. arXiv preprint arXiv:2009.03300 . · arXiv
- Dan Hendrycks, Collin Burns, Saurav Kadavath, Akul Arora, Steven Basart, Eric Tang, Dawn Song, and Jacob Steinhardt. 2021. Measuring mathematical problem solving with the math dataset. NeurIPS .
- Jordan Hoffmann, Sebastian Borgeaud, Arthur Mensch, Elena Buchatskaya, Trevor Cai, Eliza Rutherford, Diego de Las Casas, Lisa Anne Hendricks, Johannes Welbl, Aidan Clark, and 1 others. 2022. Training compute-optimal large language models. arXiv preprint arXiv:2203.15556 . · arXiv
- Wenyue Hua, Ollie Liu, Lingyao Li, Alfonso Amayuelas, Julie Chen, Lucas Jiang, Mingyu Jin, Lizhou Fan, Fei Sun, William Wang, and 1 others. 2024. Gametheoretic llm: Agent workflow for negotiation games. arXiv preprint arXiv:2411.05990 . · arXiv
- Aaron Hurst, Adam Lerer, Adam P Goucher, Adam Perelman, Aditya Ramesh, Aidan Clark, AJ Ostrow, Akila Welihinda, Alan Hayes, Alec Radford, and 1 others. 2024. Gpt-4o system card. arXiv preprint arXiv:2410.21276 . · arXiv
- Dexin Kong, Xu Yan, Ming Chen, Shuguang Han, Jufeng Chen, and Fei Huang. 2025. Fishbargain: An llm-empowered bargaining agent for online fleamarket platform sellers. arXiv preprint arXiv:2502.10406 . · arXiv
- Li-Cheng Lan, Andrew Bai, Minhao Cheng, Ruochen Wang, Cho-Jui Hsieh, and Tianyi Zhou. 2025. Exploring expert failures improves llm agent tuning. arXiv preprint arXiv:2504.13145 . · arXiv
- Roy J Lewicki and Alexander Hiam. 2011. Mastering business negotiation: a working guide to making deals and resolving conflict . John Wiley & Sons.
- Mike Lewis, Denis Yarats, Yann N Dauphin, Devi Parikh, and Dhruv Batra. 2017. Deal or no deal? end-to-end learning for negotiation dialogues. arXiv preprint arXiv:1706.05125 . · arXiv
- Aixin Liu, Bei Feng, Bing Xue, Bingxuan Wang, Bochao Wu, Chengda Lu, Chenggang Zhao, Chengqi Deng, Chenyu Zhang, Chong Ruan, and 1 others. 2024. Deepseek-v3 technical report. arXiv preprint arXiv:2412.19437 . · arXiv
- Agnieszka Mensfelt, Kostas Stathis, and Vince Trencsenyi. 2024. Autoformalizing and simulating gametheoretic scenarios using llm-augmented agents. arXiv preprint arXiv:2412.08805 . · arXiv
- John F Nash. 2024. Non-cooperative games. In The Foundations of Price Theory Vol 4 , pages 329-340. Routledge.
- David Rein, Betty Li Hou, Asa Cooper Stickland, Jackson Petty, Richard Yuanzhe Pang, Julien Dirani, Julian Michael, and Samuel R Bowman. 2024. Gpqa: A graduate-level google-proof q&a benchmark. In First Conference on Language Modeling .
- Simoni F Rohden and Diully Garcia Zeferino. 2023. Recommendation agents: an analysis of consumers' risk perceptions toward artificial intelligence. Electronic Commerce Research , 23(4):2035-2050.
- Ariel Rubinstein. 1982. Perfect equilibrium in a bargaining model. Econometrica: Journal of the Econometric Society , pages 97-109.
- Johannes Schneider, Steffi Haag, and Leona Chandra Kruse. 2024. Negotiating with llms: Prompt hacks, skill gaps, and reasoning deficits. In International Conference on Computer-Human Interaction Research and Applications , pages 238-259. Springer.
- Ryan Shea, Aymen Kallala, Xin Lucy Liu, Michael W Morris, and Zhou Yu. 2024. Ace: A llmbased negotiation coaching system. arXiv preprint arXiv:2410.01555 . · arXiv
- Jinzhu Song and Hengyu Lin. 2024. Exploring the effect of artificial intelligence intellect on consumer decision delegation: The role of trust, task objectivity, and anthropomorphism. Journal of Consumer Behaviour , 23(2):727-747.
- Michelle Vaccaro, Michael Caoson, Harang Ju, Sinan Aral, and Jared R Curhan. 2025. Advancing ai negotiations: New theory and evidence from a large-scale autonomous negotiations competition. arXiv preprint arXiv:2503.06416 . · arXiv
- Frank F Xu, Yufan Song, Boxuan Li, Yuxuan Tang, Kritanjali Jain, Mengxue Bao, Zora Z Wang, Xuhui Zhou, Zhitong Guo, Murong Cao, and 1 others. 2024. Theagentcompany: benchmarking llm agents on consequential real world tasks. arXiv preprint arXiv:2412.14161 . · arXiv
- An Yang, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chengyuan Li, Dayiheng Liu, Fei Huang, Haoran Wei, and 1 others. 2024a. Qwen2. 5 technical report. arXiv preprint arXiv:2412.15115 . · arXiv
- Shu Yang, Shenzhe Zhu, Ruoxuan Bao, Liang Liu, Yu Cheng, Lijie Hu, Mengdi Li, and Di Wang. 2024b. What makes your model a low-empathy or warmth person: Exploring the origins of personality in llms. arXiv preprint arXiv:2410.10863 . · arXiv
- Shu Yang, Shenzhe Zhu, Zeyu Wu, Keyu Wang, Junchi Yao, Junchao Wu, Lijie Hu, Mengdi Li, Derek F Wong, and Di Wang. 2025. Fraud-r1: A multi-round benchmark for assessing the robustness of llm against augmented fraud and phishing inducements. arXiv preprint arXiv:2502.12904 . · arXiv
- Junchi Yao, Jianhua Xu, Tianyu Xin, Ziyi Wang, Shenzhe Zhu, Shu Yang, and Di Wang. 2025. Is your llm-based multi-agent a reliable real-world planner? exploring fraud detection in travel planning. arXiv preprint arXiv:2505.16557 . · arXiv
- Biao Zhang, Zhongtao Liu, Colin Cherry, and Orhan Firat. 2024. When scaling meets llm finetuning: The effect of data, model and finetuning method. arXiv preprint arXiv:2402.17193 . · arXiv
- Shaokun Zhang, Ming Yin, Jieyu Zhang, Jiale Liu, Zhiguang Han, Jingyang Zhang, Beibin Li, Chi Wang, Huazheng Wang, Yiran Chen, and 1 others. 2025. Which agent causes task failures and when? on automated failure attribution of llm multi-agent systems. arXiv preprint arXiv:2505.00212 . · arXiv
- Xue Zhao, Weitao You, Ziqing Zheng, Shuhui Shi, Yinyu Lu, and Lingyun Sun. 2025. How do consumers trust and accept ai agents? an extended theoretical framework and empirical evidence. Behavioral Sciences , 15(3):337.
- Luowei Zhou, Pei Yang, Chunlin Chen, and Yang Gao. 2016. Multiagent reinforcement learning with sparse interactions by negotiation and knowledge transfer. IEEE transactions on cybernetics , 47(5):1238-1250.
Cited by
Sorted by citations
The Coasean Singularity? Demand, Supply, and Market Design with AI Agents
Shahidi, P., Rusak, G., Manning, B., Fradkin, A., & Horton, J. (2025) · doi:10.3386/w34468 · Cited by 18 · see citing works
- Authors
- Zhu, S., Sun, J., Nian, Y., South, T., Pentland, A., & Pei, J.
- Affiliations
- University of Toronto (CA)
- Google DeepMind (United Kingdom) (GB)
- Massachusetts Institute of Technology (US)
- Stanford University (US)
- Venue
- arXiv
- Publisher
- arXiv
- Published
- 2025
- ISSN
- 2331-8422 (ISSN-L)
- arXiv
- 2506.00073
- OpenAlex
- W4415249953
- Cited by
- 4
- References
- 2
- Citations by year
- 2026: 3 · 2025: 1
- Type
- posted content
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University) (cc-by)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Multi-Agent Systems and Negotiation
- Physical Sciences › Computer Science › Artificial Intelligence › AI in Service Interactions
- Social Sciences › Safety Research › Ethics and Social Impacts of AI
- Keywords
- AI agents, large language models, automated negotiation, consumer market, transaction automation, benchmarking, behavioral anomalies, financial losses, consumer protection
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
