Agentic commerce: why AI shopping agents make brands matter more

What do AI agents mean for brands?

AI agents make brands more important, because an agent cannot reroute a purchase the customer asked for by name. Agentic commerce hands searching, comparing and buying to software that acts on a person's intent. The brand that shaped that intent wins. The brand that didn't becomes a line in a comparison table.

At IW Agency, we think this is the biggest commercial question brand leaders face over the next five years. The risk is not that machines take over shopping. The risk is that machines show which brands mean something, and which never did.

Picture a weekly grocery order handed to an assistant. "Butter" gets the agent's best guess on price and availability. "Lurpak" gets Lurpak.

Conclusion: A brand name in the prompt is the strongest shelf position left.

Are AI agents really doing the shopping yet?

Not fully, but they already shape what gets bought. McKinsey's March 2026 analysis of European consumers says people "are using AI to shape what they buy, but not how they buy—at least not yet." AI, it adds, is becoming "the primary interface for discovery, comparison, and recommendation—the point in the consumer journey where preferences form and winners emerge" (McKinsey, 2026).

In the US, this behaviour is already mainstream. A report by ICSC and McKinsey found that 68 per cent of consumers had used at least one AI tool while shopping in the past three months. It also found that 62 per cent had used AI to compare brands, models, prices or reviews (Retail Dive, 2026).

The traffic is valuable too. Adobe reported that AI traffic to US retailers grew 393 per cent year over year in the first quarter of 2026. By March, revenue per visit from AI referrals was 37 per cent above non-AI traffic (Yahoo Finance, reporting Adobe, 2026). "AI is quickly becoming the primary interface between consumers and their favorite brands," wrote Vivek Pandya, director of Adobe Digital Insights.

McKinsey estimates that agentic commerce could orchestrate $3 trillion to $5 trillion globally by 2030 (Digital Commerce 360, 2025). The exact figure matters less than the direction. The moment of choice is moving into a conversation with a machine.

How do AI agents choose products?

An agent chooses from three things: what the person asked for, what it can understand about each option, and what it can justify as the best fit. We see these as three moments where a brand wins or loses. We call them the brief, the reading and the shortlist.

• The brief is the request. A named brand is a constraint the agent works within. Memory, feeling and habit put that name there, long before the conversation starts.

• The reading is how the agent understands you. It draws on your site, your product information and what others write about you. If you say what everyone in your category says, the agent repeats that.

• The shortlist is the choice when no brand is named. Here the agent needs a reason it can put into words: a clear difference, a clear use or a clear audience.

Sameness is the enemy at the reading stage. Rye bread sat in a category "long defined by muted tones and functional messaging" before Hatting set out to change it. An agent flattens a category like that into price and pack size. In e-commerce, the Nordic Nest case started from the same problem: turning a generic online shopping platform into a premium brand experience.

Conclusion: If the agent can't describe your difference, you don't have one.

Doesn't an AI agent just pick the cheapest option?

Sometimes, yes, and this is the strongest argument against investing in brand now. If agents optimise for price, specifications and delivery time, emotion looks like a cost that a machine will ignore. Why build feeling for a buyer that has none?

The answer is that the agent is not the buyer. It carries out someone's intent, and people still form that intent. McKinsey places AI's influence at the point "where preferences form", and shaping preference is what brands are built to do. A strong brand reaches the agent already decided on, as a name in the brief.

The argument does hold for brands that stand for nothing beyond their spec. Private labels already compete by imitating the leaders. Göteborgskex faced a cookie category crowded by copycats, and OLW risked losing its emotional edge as private labels multiplied. Agents will add to that pressure. In our view, the answer is more distinction, not a lower price.

Conclusion: Agents commoditise the forgettable. They follow the remembered.

What does this mean for your brand?

• Audit how AI assistants describe you. Ask the main assistants what your brand is, who it is for and how it differs. Compare their answers with your strategy. The gaps between the two give you your brief.

• Make your difference sayable. Write down, in one sentence, why someone should choose you. If the reason needs a mood board to land, an agent cannot repeat it. A clear brand platform gives the agent its source text.

• Get your brand architecture straight. Agents need to know which product belongs to which brand, and why. People tolerate portfolio confusion. In an agent's answer, it turns into errors.

• Keep investing in feeling. Being asked for by name is a position no agent can take away. Distinction and desire are built in people's heads, not in product feeds.

• Treat product data as brand expression. Machines now read and quote the names, descriptions and claims in your product listings. Write them with the same care as your packaging.

Frequently asked questions

What is agentic commerce?

Agentic commerce is a model in which AI agents shop, negotiate and transact on behalf of people. McKinsey describes it as turning shopping from separate steps (searching, browsing, comparing and buying) into a continuous flow driven by intent.

Is agentic commerce relevant in Europe yet?

Yes, at the decision stage. McKinsey's 2026 research across France, Germany and the UK finds that Europeans use AI to shape what they buy. Fully autonomous purchasing is still ahead.

Will AI agents replace brand advertising?

No. Agents act on intent that people bring to them, so building preference becomes more valuable. What changes is that a brand must also be easy for a machine to understand and describe.

How do I check how AI assistants describe my brand?

Ask them the questions your customers ask, such as "what is the best option for this need?" or "how does this brand compare with that one?" Record the answers over time and note what is wrong or missing. Then trace each error back to its source.

Chosen by people, carried out by machines

Agents will do more of the buying, but people will still decide what is worth buying. The brands that are clear, recognisable and loved get named in the brief and described correctly on the shortlist. If you want to talk about what that means for your brand, write to us at hello@iwagency.com.

Sources

• Europe's agentic commerce moment: Decision influence is here; execution is coming — McKinsey & Company, 2026

• McKinsey forecasts up to $5 trillion in agentic commerce sales by 2030 — Digital Commerce 360, 2025

• US agentic commerce revenue forecast to reach $1 trillion by 2030 — Retail Dive, 2026

• AI Traffic to US Retailers Jumps 393% in Q1 as Agentic Shoppers Outspend Humans — Yahoo Finance, 2026