Agentic commerce: The next legal frontier in AI-powered shopping

Artificial intelligence is no longer just a search tool - it is becoming a purchasing agent.  

Leading AI developers, including OpenAI, Google, Microsoft and Perplexity, are rolling out autonomous "agent" capabilities that can search for products, compare prices, and complete purchases on behalf of consumers. This evolution - known as agentic commerce - represents a profound shift in how goods and services are bought and sold. 

For example, OpenAI's Agent can navigate supermarket websites, fill shopping baskets, and return control to the user to complete payment. Microsoft's "Actions" feature, Google's AI-powered shopping mode, and Perplexity's "Comet" browser all demonstrate the same trend: agents shifting from passive information providers to active commercial actors. 

The implications extend far beyond technology. Agentic commerce disrupts traditional e-commerce channels, alters brand-consumer relationships, and challenges existing legal frameworks built on the assumption that people - not algorithms - make purchasing decisions. Businesses, regulators and consumers will all need to adapt to a world where AI agents transact on our behalf. 

Market impact

The commercial implications are enormous. If AI agents become the primary gateway to shopping:

• Consumers may bypass traditional e-commerce websites altogether.
• Search volumes on platforms like Google could fall dramatically - Gartner forecasts a 25% reduction by next year due to AI-driven discovery.
• New revenue models are emerging. OpenAI, for instance, plans to charge merchants commission for sales completed through its integrated checkout system.

This represents a new commercial ecosystem where brands compete not for clicks, but for algorithmic visibility within AI systems.

Strategic challenges for retailers and brands

As AI transforms shopping experiences, retailers must adapt to new rules of discovery, marketing and customer interaction. Brands that fail to optimise for AI ecosystems risk losing visibility, relevance and direct consumer relationships. The following are key strategic considerations for any brand selling online:

• Visibility in AI Recommendations: AI agents rely on semantic search and product metadata. For example, consumers may no longer type "red dress" but "something elegant for a summer wedding in the South of France." Retailers will need to optimise product catalogues and descriptions for natural language discovery.
• New Marketing Paradigms: Traditional digital advertising may lose effectiveness. AI systems prioritise technical factors such as site speed and clarity of text descriptions. Brands will need to experiment with content designed for bots as much as for humans.
• Disintermediation Risk: If transactions are completed entirely within AI platforms, brand websites may see reduced direct traffic. Retailers must prepare for commission-based revenue models and reduced control over the consumer interface.

Legal and regulatory challenges

Agentic commerce fundamentally disrupts existing legal frameworks designed for human decision-makers, creating urgent compliance gaps that businesses must address – for example:

• Data consent and authority - Establishing valid legal frameworks for AI agents to process personal data and make purchases without explicit human approval for each transaction.
• Contract formation liability - Determining who bears responsibility when AI agents make erroneous purchases, "hallucinate" products, or fail to properly communicate terms and condition.
• Security and fraud prevention - Protecting against AI agent manipulation (25% vulnerability rate) and allocating liability between consumers, merchants, and AI providers when systems are compromised.
• Algorithmic transparency requirements - Ensuring consumers understand how AI agents select products while balancing commercial sensitivity of recommendation algorithms under consumer protection laws.
• Cross-border regulatory compliance - Navigating inconsistent AI governance frameworks across jurisdictions where no comprehensive agentic commerce regulations yet exist.

Regulatory landscape

The pace of legal and regulatory development is failing to match the speed of artificial intelligence innovation. Agentic commerce exemplifies this gap: whilst AI agents are already transforming consumer shopping and business sales models, no comprehensive framework addresses the fundamental legal questions this technology creates.

The EU's AI Act represents the most ambitious regulatory response, but this legislation predates agentic commerce and lacks provisions for autonomous purchasing agents. Consequently, businesses navigate fragmented regulations - spanning data protection, consumer rights, contract and competition law - without coherent guidance on their application to AI-mediated transactions.

This regulatory vacuum forces businesses to operate through legal uncertainty, constructing compliance strategies from interpretations of outdated frameworks rather than purpose-built rules. Regulators will inevitably need to fundamentally reconceptualise core legal principles - agency, consent, and fair commercial practice - to govern a marketplace increasingly dominated by algorithmic rather than human decision-making.

Conclusion

Agentic commerce is not a distant prospect - it is already reshaping the commercial landscape. The technology challenges long-standing legal concepts of consent, authority, and consumer choice, while simultaneously opening vast new commercial opportunities.

Businesses that move early - adopting governance frameworks, engaging with regulators, and adapting to AI-driven distribution - will be best positioned to thrive. For lawyers and clients alike, the challenge is to evolve with the technology, ensuring that trust, fairness and compliance remain central to this next chapter of commerce.

For further reading on AI agents and Agentic AI, see our articles on The rise of AI agents, Staying Ahead in the Age of Agentic AI and Agentic AI: A Short Introduction and Key Legal Considerations.

Authors: Michelle Sally, Tom Sharpe, Alex Williamson, Emily Hinton, Sol Pearson 

This publication is intended for general guidance and represents our understanding of the relevant law and practice as at September 2025. Specific advice should be sought for specific cases. For more information see our terms & conditions.