When AI Does the Shopping, Which Brands Make the Shortlist?

Conceptual illustration of AI shopping agents selecting products, with one orange package diverted from grey alternatives on a conveyor belt.

AI shopping agents are changing how people discover, compare and choose products. For marketers, this raises a bigger question: how do brands make the shortlist, and where should the next media euro go?

Imagine planning a weekend away with your family.

You tell your AI assistant what matters: somewhere quiet, good food, activities for the children and a realistic total price. It returns three hotels, explains the differences and helps you choose.

Convenient? Absolutely.

But what happened to all the hotels you never saw? The carefully designed websites, the paid search campaigns, the inspirational videos?

Somewhere between your request and the recommendation, a shortlist formed.

For brands, understanding that moment could become one of marketing’s most important challenges.

How AI shopping agents change product discovery

On 11 November, Zalando described how its Assistant now supports product comparisons, size recommendations, basket preparation and order tracking. According to the company, customer queries in the Assistant are three times longer than traditional search queries. People can explain their situation rather than squeeze it into a few keywords.

That distinction matters.

“Waterproof jacket” describes a product category.

“I need a jacket for a wet November weekend outdoors, but I don’t want to look as though I’m climbing Everest” describes a person, a situation and a compromise.

My view is that this creates a new competitive challenge: brands need to communicate clearly enough to be considered in that context, while remaining distinctive enough for people to want them.

Who actually controls the recommendation?

We should be careful about treating AI as a neutral personal shopper.

A retailer’s assistant operates within a commercial environment. An independent assistant may draw on different sources, partnerships and technical access. A sponsored placement and an organic recommendation are different things.

The questions become: which products can the assistant access, what evidence does it use, and how clearly does it distinguish advertising from advice?

For marketers, “being visible in AI” is therefore too vague to be a strategy. Visibility needs a context: which assistant, which market, which customer need and which stage of the decision?

What this means for media spending

As AI shopping agents influence more purchase decisions, marketers need to examine which placements still reach customers and generate additional demand.

The immediate response should be to examine how media contributes to decisions.

Reuters recently reported that AI shopping agents could reduce exposure to retailer websites and advertising placements. It also reported that Amazon had seen no meaningful impact on its advertising from third-party agents so far. The direction of change deserves attention; the timing and scale remain uncertain.

That argues for disciplined experimentation.

I would put three kinds of spending under closer scrutiny:

  • Repeated exposure with little evidence of additional demand. High impression counts can conceal weak commercial impact.
  • Search activity that mainly captures customers already looking for the brand. Reported conversions deserve a closer look at what would have happened without the ad.
  • Traffic acquisition that leads to incomplete information or disappointing experiences. More visitors will rarely solve an unclear offer.

These are useful questions today. AI makes them more urgent.

Where I would invest

Marketer reviewing hotel recommendations from AI shopping agents alongside campaign photos of advertising, events and customer service.
Investing in visibility means connecting AI discovery with memorable human experiences.

Preparing for AI shopping agents means coordinating paid media with accurate product information, credible evidence and a dependable customer experience.

I would organise investment around five priorities, with paid media working alongside product, content and customer experience.

Investment priorityWhat I would fundWhat I would assess
Brand demandDistinctive creative, relevant video, social, OOH and partnershipsWhether more people remember, seek out and consider the brand
Useful product informationAccurate feeds, clear specifications, availability, pricing and practical FAQsInformation quality, discovery and conversion
Independent evidenceCredible reviews, expert coverage, customer cases and relevant PRQuality of coverage, consideration and referral activity
Customer experienceBetter advice, booking, delivery, service and recoveryRepeat purchase, complaints, returns and recommendations
Controlled experimentsRelevant AI discovery and advertising testsAdditional sales, acquisition cost and customer quality

The balance will vary. A hotel, a software provider and a consumer electronics brand should arrive at different allocations.

1. Give people a reason to ask for you

There is a meaningful difference between asking an assistant for “a good hotel” and asking it to check availability at a particular hotel.

Brand building can influence the starting point of the conversation.

I would continue investing in creative work that builds recognition and preference. The channel should follow the audience and the task. A memorable physical experience, a strong video or a relevant partnership can all contribute.

2. Make the offer easy to understand

Google has introduced additional Merchant Center attributes intended to support conversational discovery, including answers to product questions and information about compatible accessories or substitutes. That is a practical signal that the quality of product information deserves investment.

For a hotel, “family friendly” is a claim. Connecting rooms, children’s meal times, pool restrictions and transparent additional charges help someone make a decision.

Clear information helps customers and gives discovery systems more useful material to work with. It cannot guarantee a recommendation.

3. Build evidence beyond your own advertising

Independent reviews, relevant editorial coverage and specific customer stories give prospective buyers ways to check a promise.

I would fund those activities for their credibility and usefulness. Their influence on AI recommendations should be tested rather than assumed.

4. Invest in the experience people remember

My background in hospitality taught me how quickly a polished promise meets reality.

The welcome. The room. The meal. The response when something goes wrong.

Marketing can create an expectation. Operations determine whether that expectation was justified.

That is why part of the growth conversation belongs with service and product teams. A recurring complaint may point to a better investment than another acquisition campaign.

How much should move?

There is no defensible universal percentage.

For an established advertiser, I would start with a bounded 90-day test funded from activity with the weakest evidence of incremental return. Protect effective campaigns, define the hypothesis and agree on the success criteria before spending.

Keep the accounting clear, too: product data, service improvements and paid media are different budget categories. They need coordination and appropriate measures.

Alongside sales and acquisition cost, I would track AI recommendations across a consistent set of relevant customer questions. That is a diagnostic measure. Inclusion in an answer is useful only if it eventually contributes to a worthwhile business outcome.

The shortlist is only the beginning

My working thesis is that brands will need to earn consideration in more places, often before a customer reaches their website.

That makes distinctive communication, dependable information and a credible customer experience valuable investments.

Getting recommended creates an opportunity. Delivering on the recommendation creates the relationship.

AI shopping agents may help shape the shortlist, but the experience after purchase determines whether customers return.

If an assistant compared your brand tomorrow, what evidence would give it a good reason to put you on the shortlist?

Your Next Customer Might Not Be Human. But Your Brand Still Needs to Matter.

Your next customer might not be human.

10 insights marketing leaders should prepare for in the age of agentic commerce

By Andreas Zschischang | September 9, 2026

The next major shift in marketing may not be another channel, platform or content format. It may be a new participant in the customer journey: the AI agent.

10 insights to prepare the age of agentic commerce
10 insights to prepare the age of agentic commerce.

The latest Future Marketing Index, published on September 8, 2026, places Agentic AI directly at number two among current marketing trends. Half of respondents consider it relevant today, rising to 61% when looking ahead. With DMEXCO 2026 approaching under the theme “Scaling Intelligence”, the conversation is moving from experimentation toward measurable business value.

But what happens to brand preference when an AI assistant researches, compares and potentially purchases on behalf of a customer?

My view: the future is not Brand vs. AI. It is Brand + AI. Marketing must prepare for a world in which human emotion and machine-readable relevance work together.

1. The customer journey is becoming a decision network

The traditional funnel assumes that people move through a sequence of touchpoints. AI agents can compress discovery, comparison and recommendation into a single interaction.

Prepare: Map where an agent could influence your existing journey, from initial research to post-purchase support.

2. Brand preference must become machine-understandable

A strong brand creates associations in the human mind. An AI agent also needs accessible, reliable information to understand why a product is relevant.

Prepare: Translate your positioning into clear product attributes, use cases, differentiators and trustworthy content without losing the emotional brand story.

3. Product data is becoming a marketing asset

Incomplete specifications, inconsistent descriptions and outdated availability can undermine discoverability and recommendation quality.

Prepare: Treat product information management as a strategic capability shared by marketing, commerce and IT, not merely an operational database.

4. Trust will become a competitive signal

AI-assisted recommendations depend on information that can be assessed. Reviews, transparent policies, credible claims and consistent brand information may influence how products are evaluated.

Prepare: Audit the evidence behind your claims and ensure that reviews, certifications, service policies and product details are accurate and accessible.

5. The experience must work for humans and agents

Humans value inspiration, reassurance and memorable experiences. Agents require structured information, clear options and reliable processes. These needs are different, but complementary.

Prepare: Design a dual-layer experience: emotionally compelling for people and functionally understandable for machines.

6. The last mile matters more than ever

A recommendation is only valuable if the product can actually be purchased. Poor availability, unclear delivery information or a complicated checkout can break the journey.

Prepare: Connect marketing promises with inventory, pricing, fulfillment and customer service. The entire commercial experience becomes part of the brand.

7. AI visibility is not the same as business growth

Being mentioned by an AI system is not automatically a commercial success. The important question is whether AI-assisted discovery generates qualified demand, profitable sales and long-term customers.

Prepare: Develop measurement beyond rankings and mentions, including qualified traffic, conversion, incremental revenue, margin and customer lifetime value.

8. Human experiences may become more valuable

As digital discovery becomes increasingly automated, physical and personal experiences can provide differentiation that is difficult to reduce to a product comparison.

Prepare: Invest selectively in retail, events, demonstrations, expert advice and other moments where customers can experience the brand rather than simply evaluate its attributes.

9. Marketing can no longer prepare in silos

Agentic commerce connects brand strategy, content, data, technology, legal, sales and operations. A campaign team alone cannot solve the underlying readiness challenge.

Prepare: Establish a cross-functional working group with clear ownership, practical use cases and shared commercial KPIs.

10. Start with a use case, not an AI transformation slogan

The market is still evolving. Not every category will adopt autonomous purchasing at the same speed, and customers will retain different levels of control depending on the purchase.

Prepare: Select one product category or customer journey, test AI-assisted discovery and comparison, identify the gaps and build a measurable roadmap before scaling.

The strategic question

The brands that succeed will not necessarily be those producing the most AI-generated content or deploying the largest number of agents. They will be those that combine a distinctive brand with reliable data, connected commerce and a customer experience that delivers on its promise.

DMEXCO’s “Scaling Intelligence” is therefore a useful challenge for marketing leaders: move beyond the technology itself and ask where intelligence creates real customer and business value.

The next customer journey may not be entirely human. The responsibility to create a meaningful brand experience still is.

Further reading and event references

  • BVDW, Marketing-Trends 2026: Agentische KI steigt in Top 3 ein, September 8, 2026.
  • DMEXCO 2026, Scaling Intelligence, September 23–24, Cologne.
  • DMEXCO session, Agentic Commerce: Wenn KI einkauft, September 23, 2026.
  • DMEXCO keynote, The Hidden Intelligence Behind AI Assisted Commerce, September 23, 2026.

#AgenticAI #AgenticCommerce #BrandStrategy #CustomerExperience #MarketingLeadership #DMEXCO2026