Amazon Alexa is Personal - What you need to know from our AI Consumer Panel

For years, brands have focused on search. They have worked hard to improve shelf position, Amazon ranking, retail media visibility, review scores, paid keywords and comparison tables. AI shopping assistants change that.

The launch of Amazon Alexa for shopping integrates Rufus into the main shopping flow for Amazon. The first menu item is now "alexa for shopping" which opens up on the left into a conversation panel.

In the panel, the shopper talks to alexa about what they are looking for and naturally alexa makes personalised recommendations.

Alexa is clear (you just have to ask her, her answers are personal). To understand where you are, you have to understand how she answers across the breadth of shoppers not just one laptop in a head office somewhere.

What the air fryer data shows

Our recent analysis of UK air fryer questions using our AI Consumer Panel gives a clear view of how shoppers are already using AI. The questions were not neat, linear or brand-led. They were real, practical and full of context.

Shoppers asked:

• Which air fryer is best?

• Which is best for a family of four?

• Are air fryers cheap to run?

• Are they safe?

• Can they cook sausages?

• Are they healthier?

• Are they easy to clean?

• Can they replace an oven?

The biggest theme was buying advice and the best models, accounting for around 34% of the sample. But the next themes are just as important.

Cooking performance and speed accounted for 22% of the questions. General air fryer information accounted for 18%. Value, cost and energy accounted for 11% of the questions.

So the AI shopping journey is not simply: “Which product should I buy?”,  It is: “Help me understand whether this product fits my life.” That is the point FMCG teams should pay attention to.

Why this matters for FMCG

Air fryers are not an FMCG category, but the behaviour is highly relevant to food, household, personal care, health and beauty brands.

Consumers are using AI to resolve practical trade-offs:

health v taste convenience v quality price v performance family size v storage premium v value

These are the same tensions that shape supermarket baskets every day. The difference is that AI can turn these questions into an instant shortlist.

The new shelf is a shortlist

One of the strongest insights from the air fryer work is that similar questions can produce different shortlists.

Four shoppers can ask near-identical questions about the best air fryer for a family of four and receive different recommendations depending on wording and context.

One shopper may receive a general top-rated list. Another may receive value-led options. Another may receive deal-led recommendations. Another may receive a size-led answer focused on larger capacity models.

These are actual examples from our AI Consumer Panel survey:

That is the future of shopping discovery. The old digital shelf rewarded visibility against relatively stable keywords. The new AI shelf rewards brands that are eligible across many interpretations of the same need.

A brand does not only need to rank for “best air fryer”. It needs to appear when AI interprets “best” as family-friendly, affordable, energy-efficient, easy to clean, compact, safe, premium, healthy or good for batch cooking.

For FMCG, the equivalent is clear.

• “Best cereal” may mean healthiest, highest protein, lowest sugar, best for children, best value, best for gut health or best tasting.

• “Best laundry detergent” may mean sensitive skin, stain removal, sustainability, fragrance, value or cold-wash performance.

• “Best snack” may mean high protein, low calorie, lunchbox-friendly, vegan, indulgent or premium.

The commercial question is no longer just: what keywords are shoppers searching?

It is: what decision logic is AI applying when shoppers ask broad, emotional or practical questions?

“Best” is not one question

Around 32% of the air fryer questions included the word “best”, but “best” did not always mean the same thing.

Some shoppers meant best overall. Some meant best value. Some meant best for a family. Some meant best for a small kitchen. Some meant best for health. Some meant best for chips. Some meant best compared with an oven. Some meant: please make the decision for me.

“Best” becomes a bundle of inferred attributes such as reviews, popularity, capacity, price, energy use, features, reliability, brand trust and availability.

For brands, this means positioning cannot be one message only. Brands need to be clearly linked to the real needs shoppers are asking about. Brand visibility is built from attributes

Across the Amazon answers, the most frequently mentioned air fryer brands were:

1. Ninja

2. Tefal

3. Tower

4. Philips

5. Cosori

6. Instant

7. Russell Hobbs

8. Breville

AI visibility is not just about brand fame. It is also about model naming, attribute clarity, product content and how easily a brand can be retrieved against specific needs.

AI does not just recommend products. It teaches the category.

The second biggest theme in the air fryer data was cooking performance and speed.

Shoppers wanted to know whether air fryers cook faster than ovens, whether they are good for sausages, whether they can bake cakes, whether they can replace grills and how well they handle everyday meals.

That is a reminder that AI shopping is partly educational. It not only directs shoppers to brands. It helps define what the category is for. That is hugely relevant for FMCG.

Many categories depend on consumer education: how to use a product, when to use it, what it replaces, whether it is healthy, whether it is suitable for children, how it compares with alternatives and how it fits into a routine.

• If AI frames a category around speed and family convenience, certain brands benefit.

• If it frames a category around price and running costs, value brands gain ground.

• If it frames a category around health, brands with clear health, ingredient or usage credentials may be advantaged.

Brands need to influence the explanatory layer of the category - the answers AI gives before a brand is even shortlisted.

The battle is moving from persuasion to eligibility

Traditional marketing often focuses on persuasion: why my brand is better. AI commerce introduces an earlier hurdle: whether the brand is eligible to be considered at all.

Two shoppers can ask similar questions and get different answers because the assistant has interpreted their needs differently.

In the air fryer data, recommendation criteria included capacity, customer reviews, price, energy efficiency, ease of cleaning, number of drawers, cooking versatility, compactness, safety features and suitability for family size.

For senior teams, this has an important implication: AI assistants may strengthen category leaders because they have more review volume, more content, more historical visibility and broader recognition.

But they may also create openings for challenger brands that own a specific use case clearly.

A challenger does not need to win “best overall” if it can win:

• best for small kitchens

• best value best for students

• best for family batch cooking

• best sustainable option

In an AI-led shopping world, niches become recommendation triggers. One prompt is not enough

One of the most important findings is that AI answers can feel definitive even when small wording changes alter the shortlist.

That creates a measurement problem for brand owners. Asking “What is the best air fryer?” once tells you very little.

Brands need to test prompt families, shopper personas, household contexts, price sensitivities, dietary needs, usage occasions and different versions of the same question. That is where real consumer behaviour matters.

The CheckoutSmart AI Consumer Panel is valuable because it captures real questions and real answers from real people using AI tools in everyday life.

That helps brands understand what shoppers are actually asking, what AI is actually saying and which brands, retailers and sources are being surfaced. This is a different type of research asset.

What consumer brands should do next...

There are five practical actions for FMCG teams.

1. Map the real questions consumers ask, not just the keywords they search. “Is this healthy?”, “Is this good for my child?”, “What is best value?” and “What should I buy for a family of four?” are different commercial moments.

2. Audit AI visibility by use case. Do not only test whether your brand appears for “best category”. Test value, health, convenience, sustainability, family, premium, sensitive needs, occasions and substitutions.

3. Build content for decision criteria. AI needs evidence to associate brands with attributes. Reviews, specifications, product descriptions, FAQs, retailer content, expert content and third-party mentions all matter.

4. Monitor competitor adjacency. The important question is not just “are we recommended?” It is “who are we recommended alongside, and why?”

5. Treat AI as a new research channel. The questions shoppers ask AI may reveal anxieties that surveys miss, including safety worries, confusion about value, uncertainty about usage, distrust of claims or hidden household constraints.

To know more about how the CheckoutSmart AI Consumer can help you understand more about real consumer behaviour in your category, brand or topic, contact us using the form below.