UK Regulators' Joint Letter on AI Agents: What It Means

Paper-cut illustration of four small civic buildings with columns and domes joined by a dotted ribbon, a sealed paper letter, and a friendly robot with a shopping bag outside a shop window

Short answer: On 16 November 2026 the UK's four digital regulators (the CMA, FCA, ICO and Ofcom) sent a joint letter to ministers about personal AI agents. They say "a number of personal agentic products" are about to launch in the UK, and that firms bringing them to market "must understand and meet their legal and regulatory obligations, assess the risks and put appropriate safeguards in place". They want firms to "engage early and openly with the relevant regulators" before launch. The letter names three risks to deal with: fraud and scams, personal information sharing that people don't expect, and misleading recommendations. There is no new law and no new deadline. If you only use an agent inside your own business, this is a signal of where regulators are looking, not a new duty.

Last checked: 17 November 2026 (Europe/London). This is a documentation review of the DRCF's own pages. I have not spoken to any regulator, and this is not legal advice.

Paper-cut illustration of four small civic buildings with columns and domes joined by a dotted ribbon, a sealed paper letter, and a friendly robot with a shopping bag outside a shop window
Four regulators, one letter, and the shops where agents will act. Illustrative artwork; not a DRCF image.

What the regulators wrote on 16 November

The letter comes from the Digital Regulation Cooperation Forum (DRCF), the group through which the Competition and Markets Authority, the Financial Conduct Authority, the Information Commissioner's Office and Ofcom co-ordinate. The DRCF lists it under its joint statements as the "DRCF joint letter on personal AI agents". It is addressed to the Chancellor, two Secretaries of State and the Minister for AI, and it is signed by the four chief executives.

In the regulators' own order, the letter says:

  • Launches are coming. "We are aware of the forthcoming launch of a number of personal agentic products in the UK." The regulators call this "a step-change in the way AI is being used".
  • They see the upside. Personal AI agents "could be particularly valuable in helping consumers find savings on the things they need."
  • Meet your obligations before launch. Firms "must understand and meet their legal and regulatory obligations, assess the risks and put appropriate safeguards in place". The examples given are "fraud and scams, personal information sharing not in line with user expectations, and provision of misleading recommendations, decisions or information."
  • Talk to the regulators first. "We expect firms to engage early and openly with the relevant regulators before bringing their products to market." The letter adds that some firms have done this, "but engagement has not been consistently early or open across the relevant regulators."
  • Check whether you need permission. Early engagement helps firms work out which requirements apply, "including whether their activities require authorisation or registration". Later versions of a product "will also need to be tested against those requirements".
  • Users stay in charge. Agents "must give users agency over their interactions, informed by a clear and accessible understanding of agentic-specific risks", and must not "compromise the operational resilience of the wider systems with which they interact."
  • Partners have duties too. Businesses "considering commercial relationships with AI agent providers, including financial services firms and telecoms companies, should also engage with us early. Their own obligations continue to apply when customers use personal AI agents."
  • They will act. "We will not hesitate to use our powers where we reasonably suspect breaches of the law that risk harm to consumers or businesses."

What's new compared with the 30 October statement

On 30 October 2026 the same four regulators published a joint statement saying existing law already applies to personal AI agents, and that an agent's autonomy doesn't reduce the responsibility of the firm that deploys it. I explain that statement, and the six questions it asks firms to think about, in who is responsible when an AI agent acts for you.

The 16 November letter doesn't repeat that analysis. What it adds:

30 October 2026 joint statement16 November 2026 joint letter
Who it's forIndustry and the publicMinisters (published by the DRCF)
Main messageExisting law applies to personal AI agentsMeet your obligations and engage with regulators before launch
New pointsSix questions for firms; regulators will co-ordinateLaunches are imminent; engagement so far "not consistently early or open"; check authorisation or registration; keep engaging after launch; partners such as financial services and telecoms firms should engage too
New legal duty?NoNo

What it means for a UK small business

Most readers of this site are using agents, not building them. Here is how I read the letter for three common situations. This is my reading of the regulators' words, not their guidance ‡.

1. You sell an agent, or build one into a product for consumers. This is who the letter is mainly about. Before you launch, work out which regulator covers what you do (consumer law and competition sit with the CMA, payments and financial products with the FCA, personal data with the ICO, and communications and online safety with Ofcom), and whether you need any authorisation or registration. Then contact them early. Keep a record of the risks you assessed and the safeguards you chose, because the letter says later product changes will be tested against the same requirements.

2. Your customers may use their own agents with you. The letter's line on partners is aimed at firms doing deals with agent providers, and it names financial services and telecoms companies. The underlying point, that "their own obligations continue to apply when customers use personal AI agents", is worth noting for any business that sells online ‡. If an agent places an order or asks for a refund on a customer's behalf, assume your usual consumer and data protection duties still apply.

3. You use an off-the-shelf agent inside your business. The letter doesn't ask you to contact a regulator. The three named risks are still a useful checklist: could the agent be tricked into paying or sharing data (fraud), could it share personal data in a way your customers wouldn't expect, and could it give a customer a misleading answer? Our UK GDPR guide to AI agents and supplier due diligence checklist cover the practical steps.

Paper-cut shop counter where a robot assistant hands a blank card to a customer, with a fishing hook, a padlock and a signpost above, and a shopkeeper holding a clipboard with a tick
Fraud, unexpected data sharing and misleading advice: the three risks the regulators name. Illustrative artwork.

Three checks to make this week

  1. Write down what your agent can do on its own. Can it pay, send, delete or change a customer record without a person approving it? The letter's phrase is that users must keep "agency over their interactions".
  2. Test it against the three named risks. Fraud and scams, unexpected sharing of personal information, and misleading recommendations or information. Note what would stop each one.
  3. Know your regulator. If you are building or selling an agent product, identify the relevant regulator now, before launch, not after a complaint.

If your agent handles personal data, the ICO's call for evidence on agentic AI had a closing date of 20 November 2026. Our ICO AI guidance tracker follows what the ICO publishes next.

What the letter does not do

  • It doesn't create a new law, code or certificate for AI agents.
  • It doesn't set a deadline.
  • It doesn't name the products or firms that are about to launch ‡ (the letter doesn't say which they are).
  • It doesn't change the 30 October statement. Both stand.

For the security side, attackers using agents and what to lock down, see are agent skills safe?, our AI agent security checklist and our note on the NCSC's AI cyber warning for SMEs.

Frequently asked questions

What did the UK regulators say about AI agents on 16 November 2026?

The CMA, FCA, ICO and Ofcom sent a joint letter to ministers through the Digital Regulation Cooperation Forum. It says firms launching personal AI agents in the UK must understand and meet their legal and regulatory obligations, assess the risks and put safeguards in place before launch, and engage early and openly with the relevant regulators. It names fraud and scams, personal information sharing not in line with user expectations, and misleading recommendations, decisions or information as risks to address.

Is there a new UK law for AI agents?

No. The letter doesn't create a new law, rule or deadline. The regulators' 30 October 2026 joint statement already said existing legal and regulatory frameworks apply to personal AI agents, including competition, consumer and data protection law. The 16 November letter sets out what the regulators expect of firms bringing these products to market.

Do I need to contact a regulator before my business uses an AI agent?

The letter's call to engage early is aimed at firms bringing personal AI agent products to market, and at businesses considering commercial relationships with AI agent providers, such as financial services and telecoms firms. It doesn't say a small business using an off-the-shelf agent internally has to contact a regulator. This is general information, not legal advice.

Which risks do UK regulators want AI agent firms to address?

The letter gives three examples: the risk of fraud and scams, personal information sharing not in line with user expectations, and misleading recommendations, decisions or information. It also says agents must give users agency over their interactions and must not compromise the operational resilience of the wider systems they interact with.

How is the 16 November letter different from the 30 October joint statement?

The 30 October statement set out how existing law applies to personal AI agents and how the four regulators will work together. The 16 November letter is addressed to ministers. It says a number of personal agentic products are about to launch in the UK, that engagement with regulators has not been consistently early or open, and that firms should meet their obligations before launch and expect continued engagement after it.

Sources

  • DRCF, "Personal AI Agents" page, opened 17 November 2026: https://www.drcf.org.uk/personal-ai-agents
  • DRCF, Joint statements page (lists the 16 November 2026 joint letter and the 30 October 2026 joint statement), opened 17 November 2026: https://www.drcf.org.uk/publications/joint-statements
  • DRCF, "DRCF Joint Letter on Personal AI Agents sent to the Chancellor, Secretary of State for Digital, Culture, Media and Sport, Secretary of State for Business, Innovation, Science and Trade and the Minister for AI", 16 November 2026 (PDF): https://www.drcf.org.uk/siteassets/drcf/pdf-files/agentic-ai/drcf-joint-letter-on-personal-ai-agents-sent-to-the-chancellor-secretary-of-state-for-digital-culture-media-and-sport-secretary-of-state-for-business-innovation-science-and-trade-and-the-minister-for-ai.pdf
  • DRCF, "Protecting people and businesses in an agentic economy: a joint statement by DRCF members", 30 October 2026: https://www.drcf.org.uk/publications/joint-statements/joint-statement-on-protecting-people-and-businesses-in-an-agentic-economy
  • ICO, "Agentic AI call for evidence", opened 17 November 2026: https://ico.org.uk/AgenticAI

‡ legend: ‡ = not stated on an official page I opened, or my own reading rather than the source's words. Check before relying on it.

Last checked: 17 November 2026 (Europe/London). Evidence: documentation review of the DRCF and ICO pages above. I have not contacted any regulator. General information, not legal advice.

Scroll to Top