UK prepared-meal subscriptions in AI answers: September 2026

First published 29 September 2026. Updated 29 September 2026.

Across 95 buyer questions, 828 recorded answers and the same 3 answer systems, this report compares 14 researched UK prepared-meal brands. The systems were ChatGPT, Google Gemini and Google AI Overviews, observed on 29 September 2026. It is a deliberately bounded comparison, not a whole-market index. Seven brands sell a chilled, selected, recurring prepared-meal proposition and are ranked together. Seven other brands sit in different transactions, including frozen ready meals, a meal kit and a shared dinner format. They are shown separately because a household choosing a freezer stock-up or a meal kit is not making the same purchase as someone choosing a weekly chilled meal plan. The date, questions, answer count, definitions and complete distributions are available in the public measurement receipt.

The practical question is simple: when a UK buyer asks an AI system which prepared-meal subscription might suit their week, which names appear in the recorded answers, and what kinds of public sources are exposed alongside them?

The short version

• 95 questions were assessed across a fixed set of UK prepared-meal buying situations.
• 828 completed answers form the denominator for every naming percentage in this report.
• 14 researched brands are shown, split between comparable chilled peers and different-format alternatives.
• Prep Kitchen is named in 47.7% of recorded answers in the chilled-peer set.
• The full fourteen-brand median naming rate is 8.4% within this researched set.

These figures describe this dated question set and these answer systems. They do not describe sales, customer preference, revenue, quality, or the whole UK market.

What was compared

The research uses a practical definition of a direct peer: a UK consumer can select chilled, prepared meals for an ongoing routine, receive them as a recurring home delivery and reheat them with little or no cooking. That is the transaction at stake when someone searches for a workweek lunch solution, a high-protein prepared meal or a way to reduce weekday takeaway decisions.

The chilled-peer comparison contains Prep Kitchen, Frive, Eat Píng, FuelHub, SIMMER, The Good Prep and BOZU. Each was researched as a live consumer offer. This does not mean the services are identical. Their menus, delivery coverage, portioning, pricing, fitness framing and subscription mechanics can differ materially. It means they are close enough to sit in the same decision set without pretending every food-delivery brand belongs there.

The alternative-format group is equally important, but it should not be folded into the leaderboard. COOK, Field Doctor, FieldGoods and byRuby are frozen prepared-food options. Planthood remains a meal kit that requires preparation. DabbaDrop is a planned, shared Indian dinner format. Kurami combines a London fresh Meal Path with a separate nationwide frozen offer, so its national result is held outside the chilled comparison. These are real alternatives for some buyers, yet their inclusion in a combined ranking would answer a less useful question.

Chilled prepared-meal peers

The leaderboard below uses one rule throughout: a brand counts once when it is named at least once in a recorded answer. Percentages use all 828 recorded answers as the denominator and are rounded to one decimal place.

Prep KitchenPrep Kitchen

Frive

Eat Píng

FuelHub

SIMMER

The Good Prep

BOZU

The full fourteen-brand median naming rate is 8.4%. That is a descriptive midpoint within this researched set, not a category benchmark for every prepared-food provider in the country.

The gap between the leading pair and the rest of this defined group is clear in this sample. It is not, however, proof that one service is “best”, that buyers prefer it, or that its own pages caused the mention. Answers can name several brands, and an answer can omit a brand for reasons unrelated to product quality.

The counterintuitive point is the one worth keeping in view. Prep Kitchen leads this chilled set at 47.7%, but it is absent from 52.3% of the answers. A leader can be common without being universal. For operators, that makes category coverage more useful than a hunt for one magic page. The buyer’s question, the stated need, the engine and the sources surfaced around an answer all still matter.

This also explains why a carefully bounded peer group is more honest than a sprawling table of food brands. A frozen household bundle, a recipe box and a weekly selected chilled meal may all appear around “easy dinners”, but the buyer has different constraints in each case. Treating those formats as interchangeable would inflate the appearance of a broad competitive market while reducing the commercial usefulness of the result.

Different formats, shown without a rank

The table below reports naming rates for the seven researched alternatives. It is not a leaderboard. The format column is the reason these results remain separate from the chilled-peer comparison.

COOK’s 35.3% is a good example of why the separation matters. It is a substantial result in the recorded answers, but it does not turn a freezer-based prepared-meal business into a like-for-like weekly chilled subscription. Field Doctor is also a materially different option, with a diet and condition-led proposition. Operators should read these as evidence of nearby buyer alternatives, not as a reason to collapse the formats into one commercial race.

Kurami needs a further caution. Its fresh Meal Path is relevant to a London planned-meal decision; its nationwide route is frozen. The national brand-level naming figure cannot establish fresh nationwide availability. The table therefore presents the brand as an alternative, while a future London-only comparison could assess the fresh Meal Path under a separately stated definition.

Why the question set matters

Prepared-meal buying is not one question. A person may ask for a quick lunch, a protein-forward dinner, a subscription they can pause, a low-effort replacement for takeaway, an Asian flavour profile, a freezer supply, or a meal kit. The answer may change because the buyer’s situation changes.

That is why the question set was designed around prepared-meal decisions rather than a generic “best food delivery” prompt. It includes category, comparison, problem and buying situations, so the buyer context remains clear when different food formats are considered.

The same principle applies to answer systems. This report uses ChatGPT, Google Gemini and Google AI Overviews. A result in one system should not be treated as a prediction for the others. Models, source indexes, location handling and response patterns can change. The measurement date is therefore part of the finding, not decoration.

For a plain-language account of why a defined question set, answer denominator and source record matter, see Schmitdy’s AI Search measurement methodology. The method separates a brand being named from sources shown around an answer and from commercial outcomes.

Sources surfaced alongside answers

The answer systems exposed a mix of operator domains, editorial publishers and social platforms. The counts below show recorded answers retrieving the source within the 828-answer population. They are not proof that a source caused a particular brand to be named, and a visible source is not automatically an endorsement.

There are three practical takeaways from this table.

First, the two leading chilled-peer domains and BBC Good Food are distinct source types. First-party explainers, competitor comparison material and independent editorial reviews can coexist in the same answer environment. An operator should not assume that an attractive homepage replaces a usable buying guide, or that an editorial mention replaces accurate subscription information on its own site. The inspected Good Food guide starts with the practical problem of being short on time and explains its reviewed category frame. That is a concrete reason to develop an evidence-led editorial story around a buyer problem, with no assumption of coverage or attempt at mass outreach.

Second, source-use counts are not causality. A system can expose a source without making its internal selection process legible, and several sources can sit beside the same answer. The appropriate response is to inspect the actual pages, ensure commercial facts are current, and publish useful information for a buyer rather than attempting to reverse-engineer a single answer.

Third, smaller platform counts should lead to restraint. Reddit, Facebook, YouTube and LinkedIn can matter in individual situations, but this report does not show a basis for a blanket social programme. Community participation should be useful, permitted and transparent. A brand should never manufacture recommendations, plant testimonials or treat an old thread as a distribution channel.

What a prepared-meal operator can do next

The sensible first moves are operational and editorial rather than theatrical.

Make the weekly transaction easy to understand. A buyer deciding between meal kit, takeaway and prepared meal needs clear answers on preparation, delivery day, ordering cut-off, storage, service area, minimum order, recurring subscription and pause or cancellation. These details should appear close to a plan or menu decision, with a linked explanation that stays aligned with current checkout rules. The point is not to add more claims. It is to prevent a buyer from having to infer the practical shape of the service.

Turn menu proof into decision pages. Prepared-meal operators often already hold their strongest evidence in product data: dish names, cuisines, allergens, protein figures, portion information and weekly availability. Useful collections can make that evidence easier to compare, such as a current high-protein selection with an explicit threshold, or a workweek lunch collection that says what is included and when it is delivered. The collection must use the operator’s published data accurately and avoid broad nutrition or medical promises.

Publish comparisons with honest boundaries. A page explaining prepared meals versus meal kits, freezer meals or takeaway can be genuinely useful when it describes the mechanics rather than declaring a winner. It should say when cooking is required, whether meals are chilled or frozen, how delivery works, what a buyer selects, and how recurring orders are managed. This is especially valuable in a category where neighbouring formats are common in answers but do not serve the same purchase occasion.

Operators that want to make their own reporting inspectable can also use Schmitdy’s measurement explainer and its restaurant information-page example. The latter is a hospitality example, not a prepared-meal service classification. The underlying principle transfers: start with buyer questions and accurate public facts, then retain enough evidence to distinguish an observation from an assumption.

Questions operators may have

Does 47.7% mean Prep Kitchen is the UK market leader?

No. It means Prep Kitchen was named in 47.7% of the 828 recorded answers in this dated, defined chilled-peer comparison. It does not establish whole-market leadership, sales, revenue, customer preference or product quality.

Why are COOK and Field Doctor not in the chilled-peer table?

They are substantial buyer alternatives, which is why their naming rates are reported. Their frozen and, for Field Doctor, condition-led transaction differs from a selected weekly chilled prepared-meal subscription. Combining them into one table would obscure the buyer choice the comparison is intended to examine.

What does the 8.4% median tell us?

It is the median naming rate across the fourteen researched brands. It offers a midpoint for this defined research set only. It is not a market average and should not be used to assess an unresearched provider.

Limits

This is directional evidence from a defined question set, not a definitive assessment of the market. It can answer these bounded questions:

• Which of the researched chilled services appeared most often in these answers?
• Which other meal formats entered the same buying questions?
• Which inspected sources were retrieved alongside those answers?

The practical limits are equally specific:

• The comparison is bounded to 95 questions, 828 recorded answers, three named answer systems and one measurement date.
• The fourteen researched brands are not a census of UK prepared-food operators.
• The chilled-peer leaderboard covers seven comparable chilled offers. The seven alternatives are reported separately because their formats differ.
• Answer text and source-use observations do not measure revenue, customer preference, quality, conversion or causality.
• Naming rates may change when questions, systems, location conditions, source indexes or time change.
• A source appearing beside an answer is not proof of endorsement, retrieval mechanism, commercial effect or permission to participate in the source’s community.
• COOK was detected conservatively through its uppercase brand name or official domain. A brand named with unconventional capitalisation or without a recognisable domain may be missed, so the report does not claim complete brand detection beyond the defined research set.

Use the results by buyer fit: inspect the exact answer and its sources, then improve the weakest factual explanation. Chasing one blended leaderboard score is not a growth plan.

Sources and method

• Public measurement receipt: dated population, definitions, complete brand distribution and source counts.
• Prep Kitchen: official current product and ordering information used to classify this research subject.
• Frive: official current product and ordering information used to classify this research subject.
• Eat Píng: official current product and ordering information used to classify this research subject.
• FuelHub: official current product and ordering information used to classify this research subject.
• SIMMER: official current product and ordering information used to classify this research subject.
• The Good Prep: official current product and ordering information used to classify this research subject.
• BOZU: official current product and ordering information used to classify this research subject.
• COOK: official current product and ordering information used to classify this research subject.
• Field Doctor: official current product and ordering information used to classify this research subject.
• FieldGoods: official current product and ordering information used to classify this research subject.
• byRuby: official current product and ordering information used to classify this research subject.
• Planthood: official current product and ordering information used to classify this research subject.
• DabbaDrop: official current product and ordering information used to classify this research subject.
• Kurami: official current product and ordering information used to classify this research subject.

Inspected editorial examples: Good Food’s meal-delivery review and Glamour’s Frive review. The complete measured distribution, definitions and source record are available in the public measurement receipt.

If you run a UK prepared-meal service and want to understand which buying questions your public evidence can answer, talk to Schmitdy.