CMA Focus On Agentic AI And Impact On Consumers

The CMA has recently turned its attention to agentic AI and its impact on consumers. On 9 March 2026 the CMA published a research and analysis report on Agentic AI and consumers and separate guidance for businesses on complying with consumer law when using AI agents.

The CMA's report and guidance make it clear that the consumer protection provisions and the CMA's new direct enforcement powers as set out in the Digital Markets Competition and Consumers Act (DMCC Act) apply to conduct by AI agents in the same way as they do to conduct that is generated by humans.

The latest guidance is an important reminder for businesses who use AI agents to make sure their AI agents are designed and trained to comply with the relevant consumer protection rules. Failure to do so could result in fines of up to 10% of global annual turnover for the companies involved. Now is the time to carry out a careful audit and make the necessary changes as publication of the CMA's guidance signals this has very much become an area of focus under the new consumer protection regime.

Background

Definitions of Agentic AI vary but broadly speaking it refers to autonomous systems that act as agents that are able to reason, plan, and execute complex tasks to achieve specific goals, with minimal human oversight. This is in contrast with conventional AI systems, which provide specific outputs/create content, bases on inputs and prompts. 

The CMA recognises the benefits of agentic AI and AI agents, which has the potential to substantially increase productivity and efficiency for businesses and to improve consumers' daily lives by saving time and effort thanks to systems that anticipate their needs and execute transactions on their behalf.

At the same time there are also considerable risks. The technology is at an early stage and greater autonomy for agents increases the consequences of errors. Personalised support may tip over into manipulation, bias, lock-in and loss of agency. The UK's consumer protection and competition enforcement regimes, designed to ensure consumer protection, fair play and innovation provide a strong framework for addressing these challenges as they apply regardless of whether consumers or businesses interact with people or with AI systems.

Guidance on complying with consumer law when using AI agents 

The Guidance on complying with consumer law when using AI agents (the Guidance) makes it clear that businesses are required under consumer protection legislation to treat consumers fairly, regardless of whether they interact through a person or through an AI agent. They are responsible for the conduct by an AI agent in the same way they are responsible for the conduct of an employee. This is the case even if the AI agent was designed or provided by a third party on their behalf.

The Guidance sets out the key principles to assist businesses who use AI agents to build trust and comply with the relevant consumer protection laws:

Tell customers if you use an AI agent

Businesses should be clear and open about how they use AI agents in order to build trust. Consumer protection legislation requires businesses to provide consumers with the necessary information to allow them to make informed decisions. Telling customers upfront that you use an AI agent will avoid the risk of misleading customers into believing that the service is provided by a real person.

Train AI agents to comply with consumer law

Start by considering what your AI agent will be set up to do and how that may affect your customers. Make sure the AI agent will be prompted to act in a way that respects your customers' statutory rights and terms of contracts, avoids misleading customers and to obtain all necessary consents required under consumer law. Testing and evaluating performance are also an important part of this training, and remember that you will be responsible for any breaches of the law by your AI agent.

Monitor how AI agents are performing

Businesses should regularly check how well their AI agents are performing and whether they are delivering the right results, behaving as intended and complying with consumer legislation. Human oversight is necessary to spot any errors and to ensure that AI agents are completing tasks in compliance with the relevant laws.

Refine the AI agent quickly if there is a problem

If an AI agent is not performing as expected and this may result in infringements or potential infringements of the law, businesses should act quickly to address the problem. This will be the case in particular where an AI agent interacts with large numbers of consumers, especially vulnerable consumers. 

The Guidance also contains some worked examples illustrating how AI agents should operate in a number of use cases:

Running a marketing campaign

Businesses designing a marketing campaign using an AI agent should make sure that the relevant consumer rights are identified upfront and complied with. By way of example:

• any marketing materials which contain pricing information should provide accurate information about pricing, in line with the CMA's price transparency guidance
• where incentivised consumer reviews are included in the materials these should be clearly labelled as such, in line with the CMA's guidance on fake or misleading reviews
• any special offers or discounts should be genuine and not mislead consumers

The campaign should regularly be reviewed by a human with the relevant expertise and promptly updated where necessary.

Processing refund requests

Where AI agents are used for processing refund requests, businesses should make sure that the agent is designed to take into account the relevant contractual and legal requirements, including the Consumer Rights Act 2015 and the Consumer Contracts (Information, Cancellation and Additional Charges) Regulations 2013. The AI agent's exchanges with consumers should be reviewed by a human on a regular basis to ensure that the agent's decisions accurately reflect the nature and the amount of the refund requests.

Responding to customer service queries

Designing and training an AI agent to respond to customer queries will require the AI agent to be in a position to respond accurately to queries about prices, products and the consumer's rights, give consumers the necessary information to make informed decisions and not make it difficult for consumers to exercise their rights. Someone within the business with the relevant experience should review the performance of the AI agent and how it responds to different types of queries on a regular basis.

Providing a price comparison service to customers

When using an AI agent for a price comparison service businesses should make sure that the results are accurate and that important information, such as market coverage, data searched, any limitations, ranking of results, links with suppliers are all clearly disclosed.

Comment

Although the use of agentic AI or algorithms is not necessarily problematic, businesses need to be aware of the potential compliance risks these tools can create. The CMA makes it clear that companies are responsible for putting in place the necessary safeguards to prevent infringements as a result of using these tools, both under the consumer protection and under the competition regimes, as the risk ultimately remains with them. This means carrying out a detailed risk assessment and designing compliance frameworks tailored to their specific activities.

It is also worth noting that the CMA itself is using these advanced technologies "to detect consumer and competition issues at unprecedented pace and scale". It is using agentic AI to identify potential infringements of consumer protection law across the economy to better understand consumers' experience and target its enforcement activity where it will have the greatest impact (see the CMA's blogpost on 'AI and collusion: frontiers, opportunities and challenges').

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.