Home
Anthropic's experimental AI Claude completes negotiations and transactions in e-commerce test
As artificial intelligence advances rapidly, Anthropic quietly rolled out an internal experiment called "Project Deal" last Friday, showcasing AI's potential in e-commerce. The experiment had its AI model Claude autonomously handle buying, selling, and price negotiations within a closed market environment, involving real financial transactions.

At the heart of the experiment was an internal marketplace built on Slack, where Claude took on the role of negotiator for both buyers and sellers. It first interviewed 69 employees to gather their buying and selling intentions along with personalized instructions, then carried out the negotiations autonomously. Over the course of the experiment, Claude successfully completed 186 transactions totaling over $4,000, demonstrating AI's ability to operate in real-world market settings.
To ensure thoroughness, Anthropic also created four parallel markets to compare how different AI models performed in negotiations. Results revealed that variations in model capabilities had a significant impact on transaction outcomes. For instance, when the Opus model faced off against the Haiku model, Opus clearly outperformed Haiku in negotiations, while the surveyed employees remained unaware of this difference. This insight provides a fresh perspective on future commercial interactions between AI systems.
The unveiling of "Project Deal" has drawn considerable market attention, especially over its potential threat to traditional e-commerce platforms. Following the news, eBay's stock price dropped approximately 4.5% by market close. Observers widely attribute this decline to the application of AI technology and its disruptive impact on e-commerce business models.
Although Anthropic opted to release the experiment during a period of significant market fragmentation, the market reaction was nonetheless notable, underscoring AI's disruptive potential in e-commerce.
Related article
Reliance unveils $110B AI investment plan as India accelerates tech drive
Mukesh Ambani, the billionaire chairman of India's Reliance conglomerate, announced on Thursday a ₹10 trillion (roughly $110 billion) plan to build AI computing infrastructure across India over the next seven years.Speaking at the India AI Impact Sum
Zhiyuan WITA Ends 'Naked' Robot Interaction with First Compliance Filing
The embodied intelligence sector has reached a significant milestone. According to the latest announcement from the Shanghai Cyberspace Administration, the WITA large model developed by Zhiyuan has successfully completed the filing process, becoming
Anthropic Study Links Polished AI Content to Reduced Human Thinking
When you see AI instantly produce a well-structured, logically clear piece of code or document, are you tempted to trust it without a second thought? According to AIbase, the leading AI company Anthropic recently published a research report titled "A
Related Special Topic Recommendations
Comments (0)
0/500
As artificial intelligence advances rapidly, Anthropic quietly rolled out an internal experiment called "Project Deal" last Friday, showcasing AI's potential in e-commerce. The experiment had its AI model Claude autonomously handle buying, selling, and price negotiations within a closed market environment, involving real financial transactions.

At the heart of the experiment was an internal marketplace built on Slack, where Claude took on the role of negotiator for both buyers and sellers. It first interviewed 69 employees to gather their buying and selling intentions along with personalized instructions, then carried out the negotiations autonomously. Over the course of the experiment, Claude successfully completed 186 transactions totaling over $4,000, demonstrating AI's ability to operate in real-world market settings.
To ensure thoroughness, Anthropic also created four parallel markets to compare how different AI models performed in negotiations. Results revealed that variations in model capabilities had a significant impact on transaction outcomes. For instance, when the Opus model faced off against the Haiku model, Opus clearly outperformed Haiku in negotiations, while the surveyed employees remained unaware of this difference. This insight provides a fresh perspective on future commercial interactions between AI systems.
The unveiling of "Project Deal" has drawn considerable market attention, especially over its potential threat to traditional e-commerce platforms. Following the news, eBay's stock price dropped approximately 4.5% by market close. Observers widely attribute this decline to the application of AI technology and its disruptive impact on e-commerce business models.
Although Anthropic opted to release the experiment during a period of significant market fragmentation, the market reaction was nonetheless notable, underscoring AI's disruptive potential in e-commerce.
Reliance unveils $110B AI investment plan as India accelerates tech drive
Mukesh Ambani, the billionaire chairman of India's Reliance conglomerate, announced on Thursday a ₹10 trillion (roughly $110 billion) plan to build AI computing infrastructure across India over the next seven years.Speaking at the India AI Impact Sum
Zhiyuan WITA Ends 'Naked' Robot Interaction with First Compliance Filing
The embodied intelligence sector has reached a significant milestone. According to the latest announcement from the Shanghai Cyberspace Administration, the WITA large model developed by Zhiyuan has successfully completed the filing process, becoming
Anthropic Study Links Polished AI Content to Reduced Human Thinking
When you see AI instantly produce a well-structured, logically clear piece of code or document, are you tempted to trust it without a second thought? According to AIbase, the leading AI company Anthropic recently published a research report titled "A











