Summary

  • The CMA’s revised proposal for a Google search-choice requirement closed for consultation on 9 October; no final requirement has been announced.
  • Google would decide which rival providers qualify, under published criteria and a six-month application process, then report decisions and detailed rejection reasons to the CMA.
  • AI assistants could qualify in theory, but only if they meet the criteria. The proposal does not establish that any assistant has applied or been admitted.

The difficult decision in a choice screen comes before a user sees it: which services are allowed onto the list? The Competition and Markets Authority’s revised UK proposal answers that question by assigning Google the first assessment of rival search providers. The CMA would set the proposed criteria and monitor the process, but Google would run applications and make individual eligibility decisions.

The CMA published the revised draft on 23 September after a January consultation and further evidence gathering. Its consultation closed at 5 p.m. on 9 October. The CMA said in September that it expected a final decision by year-end. That remains a stated timetable, not an adopted conduct requirement. The current case page lists the further consultation but no later decision.

The draft criteria are broad enough to include changes in how people search. A service would need to address the full range of general-search use cases, including web search; provide a functioning general-search experience at the relevant access points; and be used by a significant proportion of UK users. The CMA does not set a numerical threshold for the last test. It also says that an AI component alone should not exclude a service: an AI assistant could in theory qualify if it meets the criteria. That is not a finding that any named assistant qualifies.

Google would open an eligibility round every six months, give at least two weeks’ notice before applications open, and decide each application within two months from the start of the window. It would send written decisions to applicants and report all decisions to the CMA, including detailed reasons for rejections. The CMA says these reports would let it monitor how Google applies the criteria. The proposal therefore moves the operational admission decision to the company whose search service the rule is meant to constrain, while placing the process inside a regulator-defined and reported framework.

The interface rules matter, but they do not remove that gate. If there are 12 or fewer eligible providers, the screen would show all of them. If there are more, Google would select 12 fairly, while keeping the rest accessible. Their order would be randomized, and users would have to scroll through the list before confirming. The draft also includes first-use screens at specified Chrome and Android access points, limited follow-up prompts and a short trial option. These are proposed duties, not current user rights.

The evidence behind the screen design has a boundary. The CMA’s commissioned research comprised 40 self-guided and 40 guided qualitative tests. Participants often skimmed information and did not consistently notice that more providers appeared below the fold. The researchers describe an exploratory, small, non-random sample; it cannot show how all UK users would behave or predict market-wide switching.

The commissioning distinction is important. A March 2026 BTW article covered Google’s earlier strategic-market designation and the prospect of future interventions. This update concerns the later, revised proposal and the admission process that would decide which rivals could appear. It reports neither a final rule nor a change in Google’s market position.

Primary sources