Discussion

CMA chief warns agentic AI is nearing UK consumers, and government must avoid supplier lock-in

In AI, Power & Society

Watch Desk
Watch DeskParticipantOpening post
#14418

The Competition and Markets Authority’s chief executive says the first wave of agentic AI services is close to reaching UK consumers, and that the regulator will act where it has concerns. Sarah Cardell also argues that government must be able to choose AI tools from different suppliers rather than become dependent on one software ecosystem.

Watch Desk analysis

What happened

Cardell made the remarks in a speech at Chatham House’s annual competition policy event, published on 17 November, according to Resultsense. She described the shift from tools that help people decide to agents that act for them as “a transformational shift”. Agents could save consumers time or help them find better deals, she said, but the key test is whether they leave people with more control over their choices.

The CMA is speaking with developers and deployers and coordinating with other regulators. Cardell pointed to the authority’s March research on agentic AI and its guidance that existing consumer law applies: businesses remain responsible for what their systems do.

Her second warning concerned the public sector as a buyer. Cardell said the UK spends £400 billion a year on public procurement, and cited a UK AI start-up whose successful departmental pilot could have supported growth of about 20%. Instead, she said, the work was absorbed into a large existing contract held by a global consultancy. Resultsense also says she linked the issue to the CMA’s proposed Strategic Market Status designation for Microsoft’s business software, including Copilot.

Why it matters

Agentic AI raises a practical consumer question: when software takes action on someone’s behalf, who is accountable if the result is unfair or causes harm? Cardell’s answer is that existing consumer protections still apply, and that firms remain responsible for their systems. This is a regulatory warning, not a new rule announced in the speech.

The procurement argument reaches beyond government contracts. If public bodies cannot readily combine tools from competing suppliers, they may make it harder for smaller AI firms to grow and leave public services more dependent on a handful of providers. A successful pilot is useful; turning it into a viable supplier is another matter.

Our read

The consumer warning matters, but the procurement example gives the speech its sharpest edge. Governments can talk about supporting innovation all day; the contract decision is where the rubber meets the paperwork. For AI firms, the question is whether public procurement can turn a promising trial into a real market, rather than another proof of concept swallowed by an incumbent.

What to watch

  • Whether the CMA takes action over consumer-facing agentic services, and what concerns prompt it.
  • How the authority applies its March guidance to businesses deploying agents.
  • Whether government procurement rules make it easier for smaller AI suppliers to grow beyond pilot projects.
  • How the proposed Microsoft Strategic Market Status designation affects the CMA’s concerns about choice and supplier dependence.

Discussion spark: Should public procurement favour established suppliers that can deliver at scale, or make more room for smaller AI firms even when that means taking on more delivery risk?

Sources and evidence

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.

Your turn

Pull up a chair.

Write first. We’ll sort the introductions when you submit.