Executive Summary
A quiet shift is already underway across Europe. AI tools are starting to influence what people buy, even when checkout still happens in the usual places. McKinsey describes it plainly: decision influence is here, execution is coming. (McKinsey & Company)
That detail changes the game for mid-level CPG managers. Brand growth has always depended on being easy to find, easy to trust, and easy to choose. Agentic commerce adds a new gatekeeper: the AI assistant that narrows options, explains trade-offs, and decides which three products even make it onto the shortlist.
BCG’s analysis of European traffic patterns makes the urgency hard to ignore. AI-driven search and referral is growing quickly and is projected to reach 25% of organic visits by the end of 2026, then surpass organic visits in 2028. (BCG Global)
This post explains what agentic commerce really means for CPG, why “better marketing” is no longer enough on its own, and how mid-level leaders can build a practical capability that restores momentum in both the company and their own career. Real-world examples from Europe, the US, and Asia show what is working, what is failing, and what to build next. (McKinsey & Company)
Prefer to Listen?
The new battleground is not checkout. It’s the shortlist.
People are using AI to shape what they buy, but many are not yet letting AI complete the purchase. (McKinsey & Company)
McKinsey’s point about Europe is sharper than it first appears. That creates a transition period where many CPG teams feel safe because conversion still happens on retailer sites, in store, or in familiar e-commerce flows.
The real change sits earlier. AI is starting to filter the market before a shopper even reaches a retailer search bar. That means brands can lose without ever seeing the lost shopper. Nobody abandons your PDP, because they never reach it. Nobody compares your claims to a competitor’s claims, because the assistant filtered you out before comparison even began.
BCG’s traffic insight is the most useful “wake-up” statistic because it connects behaviour to business impact. AI-driven search visits in Europe are rising fast, with BCG projecting a jump to 25% of organic visits by end of 2026. (BCG Global) Even if that projection lands a little high or a little low, the direction still forces a new question for every brand plan.
What would need to be true for an AI assistant to recommend your brand, with confidence, to a shopper who never visits your website?
That question sounds abstract until you see what retailers and platforms are building.
Agentic commerce is arriving in layers, and that matters for your plan
Google is pushing an “agentic shopping era” narrative and has announced tools and an open standard approach aimed at helping retailers connect with high-intent shoppers using AI. (blog.google) Google’s ecosystem approach is moving toward letting shoppers discover and buy within AI surfaces.
The Verge reported that Google partnered with Gap to enable Gemini to buy clothing directly through the chatbot using Google Pay, powered by Google’s Universal Commerce Protocol (UCP). (The Verge) The details are fashion, yet the lesson is broader. A protocol layer is emerging that allows agents to complete commerce actions. Brands and retailers will increasingly be judged by how well their product data and fulfilment can be consumed by those agents.
Walmart’s experience shows the other side. Wired reported that Walmart and OpenAI’s earlier chatbot checkout experiment underperformed, partly because the design made shoppers buy one item at a time, which did not match how people shop. (WIRED) The story is useful because it punctures the hype. Agentic commerce is not magic, and early versions can fail badly when they don’t fit real behaviour.
That failure is still good news for CPG leaders who want an advantage, because it reveals what the next wave will focus on: bundling, baskets, substitutions, availability, loyalty, and post-purchase help. Those are not marketing problems. Those are system problems, which is exactly where mid-level leaders can become visible and valuable.
Agentic commerce is also global, not US-only. Rakuten announced an AI concierge on Rakuten Ichiba that uses conversational shopping to understand needs and match products, which shows how Asian marketplaces are already turning discovery into an assistant-led experience. (Rakuten Group, Inc.) Grocery retail is moving too. Grocery Dive covered Albertsons’ agentic tool aimed at simplifying online shopping, which signals grocery is not waiting for “perfect tech” before experimenting. (Grocery Dive) The Grocer reported Ocado rolling out an agentic AI chatbot to handle customer queries, starting with a limited segment of users. (The Grocer)
Put those examples together and the pattern becomes clear. Agentic commerce is not one event. It’s a set of capability releases that move decision influence earlier and earlier.
That makes this a momentum topic, not a tech topic.
Why mid-level CPG managers should care more than anyone else
Agentic commerce changes where credibility lives. Brand trust used to be shaped by advertising, shelf presence, and word of mouth, then reinforced by product experience. AI assistants introduce a new mediator that rewards proof, consistency, and clear trade-offs.
Mid-level managers often sit in the exact roles that can fix or fuel this. Innovation leaders write briefs. Insights leaders define what matters. Brand managers choose claims language. Ecommerce teams govern product content. Shopper teams translate retail reality. None of those roles owns the entire system, yet every one of them shapes whether a brand becomes “assistant-friendly.”
Gartner’s prediction for marketers adds a timing signal worth taking seriously. Gartner expects that by 2028, 60% of brands will use agentic AI to deliver streamlined one-to-one interactions. (Gartner) That means competitors are not waiting. Many are already piloting.
The career implication is obvious once you say it out loud. The manager who can translate agentic commerce into practical work becomes the person who reduces risk and creates new growth paths. That is exactly the kind of work that gets noticed.
The new funnel is “assistant-first,” and it rewards different inputs
Classic funnels assumed the consumer explored the market directly. They searched, browsed, compared, then chose. Agentic commerce creates a “delegated discovery” funnel where the shopper asks an assistant to do the heavy lifting.
That delegated discovery funnel relies on signals. Some are technical, such as structured product data, availability, and compatibility with protocols and retailer feeds. Some are behavioural, such as reviews, returns, repeat purchase, and loyalty patterns. Some are narrative, such as clear claims and credible proof.
Walmart’s early failure with OpenAI checkout shows what happens when the funnel doesn’t align with real shopping behaviour. (WIRED) The assistant did not create a natural basket experience, which reduced conversion. That detail matters for CPG because basket behaviour is where brands win share. If the agent cannot build baskets well, the entire experience suffers.
So the opportunity becomes straightforward. Brands that supply the right signals will be easier for agents to recommend, easier for shoppers to accept, and easier for retailers to support.
That is the reason agentic commerce can protect margin. Shortlists compress choice. Compressed choice reduces the need for endless discounting, because the brand competes inside a smaller set where proof and preference matter more than sheer visibility.
What it takes to “earn the AI shortlist”
Agentic commerce will not reward brands that simply talk better. It will reward brands that are easier to evaluate.
McKinsey’s framing about AI narrowing the field to a small set of options captures the point. (McKinsey & Company) Shortlists are built on reasons, not vibes. AI assistants need reasons they can state confidently. Shoppers need reasons they can believe quickly.
Three capability areas separate brands that will show up from brands that will be filtered out.
First comes product truth that is consistent across the ecosystem. Retailers, marketplaces, DTC sites, and syndicated content often drift. One description says “suitable for sensitive skin,” another says “dermatologically tested,” another says nothing. Agents will see that drift as uncertainty. Shoppers will experience it as mistrust.
Second comes proof that is easy to retrieve. An agent that can’t find credible evidence will default to safer choices. Euro-style regulation and consumer scepticism have already raised the bar on what counts as believable. Your brand should be the easiest one to justify.
Third comes a clear decision structure that fits how people choose. Agents will increasingly mediate trade-offs, especially when budgets are tight. Brands need to make those trade-offs easy to articulate.
That is where many teams lose momentum, because “clear decision structure” sounds like work that belongs to someone else.
It belongs to you, because it is insight work translated into buying logic.
Case study lessons from what’s working and what’s failing
Walmart’s OpenAI checkout underperformance is a gift, because it explains what agents must handle to become mainstream: natural baskets, bundled shopping, and retailer-native workflows. (WIRED) That will improve. When it does, brands will face a sharper form of competition because the assistant will take more ownership of choice, not just research.
Google’s ecosystem push, including its agentic commerce tools and protocols, signals a different risk. (blog.google) Brands and retailers could lose some control over discovery as the assistant layer becomes the main interface. That makes product data readiness and content governance non-negotiable.
Rakuten’s AI concierge signals another pattern. Marketplaces are embedding AI assistants inside ecosystems that already have payments, loyalty, and fulfilment. (Rakuten Group, Inc.) Brands in those ecosystems will be judged not just on product quality but on how well they help the system recommend, substitute, and explain.
Ocado’s agentic chatbot shows the retail service angle. (The Grocer) Agents will increasingly handle questions about allergens, usage, delivery, substitutions, and returns. Those interactions shape trust, and trust shapes repeat purchase.
Albertsons experimenting with an agentic tool reinforces the point that grocery is moving earlier than many CPG leaders assume. (Grocery Dive) Grocery is where baskets are built. Basket logic is where brands win. That is why CPG teams should treat agentic commerce as a core growth topic, not a novelty.
Turning agentic commerce into a practical operating model using CATSIGHT™
This topic becomes actionable when it stops being “AI commerce” and starts being “decision influence.”
A strong starting point is to treat the AI assistant as a new kind of consumer in your insight process. That consumer is not emotional in the same way a human is. It is rational, proof-seeking, and consistency-seeking. That is not a threat. It is a forcing function that improves discipline.
CATSIGHT™ makes this practical because it begins with clarity. Category and brand definition has to be explicit, because assistants rely on boundaries. “Premium” means nothing unless the assistant can translate it into attributes, proof, and trade-offs. Aim is the next anchor, because delegated discovery exists to achieve an aim faster. A shopper asks an assistant to avoid confusion, avoid regret, avoid waste, or avoid being misled. Those are aims. Those aims should shape how the brand presents choice.
Intimacy still matters, even with AI intermediating, because assistants will reflect what people ask. The language consumers use when they are tired, sceptical, or overwhelmed becomes the language the assistant will hear. That language is the bridge to being recommended. Teams that collect those phrases across markets will build better prompts, better product content, and better decision logic.
Human truth still sits underneath everything. AI might shortlist, but humans still consume, judge, and talk. The human tension remains the engine of preference. Agentic commerce changes the interface, not the underlying psychology. That is why the best brands will combine assistant-friendly proof with human-friendly meaning.
Action is the step most teams skip. Action here means building a “shortlist readiness” workstream with clear ownership and a cadence that ships improvements, not decks.
QC2™ makes it stick because this is not a marketing project
Agentic commerce touches company, consumer, brands, and process all at once. That is why many organisations struggle. Marketing may want to move quickly, legal may want to slow down, ecommerce may be understaffed, retailer teams may be measured on different outcomes, and insights may not own product content.
QC2™ is useful because it forces alignment around the system rather than around one channel. Company choices determine who owns product truth. Consumer choices determine what “proof” and “control” look like. Brand choices determine what trade-offs are acceptable. Process choices determine whether content stays consistent across markets, retailers, and platforms.
That is where momentum comes from. A mid-level leader who can convene those pieces and turn them into a repeatable rhythm becomes indispensable.
Gartner’s agentic AI prediction for brands reinforces that this is becoming mainstream. (Gartner) Waiting for the dust to settle is no longer a safe plan. The teams that build capability early tend to shape standards and secure better retailer partnerships.
The “shortlist readiness” blueprint for the next 90 days
A workable approach is to focus on three tangible outputs that change how your brand shows up in assistant-led discovery.
First output is a product truth file that is consistent and structured. That means cleaning up contradictions across retailer listings, DTC content, and syndication. Agents will reward consistency, because consistency signals reliability.
Second output is a proof pack that an assistant can cite. Proof does not mean endless scientific claims. Proof means clear substantiation for what matters in your category. The aim is to make your product the easiest to justify. That protects premium pricing.
Third output is decision logic written in human language. Assistants will increasingly compare trade-offs. Your brand should make trade-offs easy to explain, such as “best for sensitive skin when fragrance triggers irritation,” or “best for weeknight speed when time is the enemy,” with clear supporting evidence.
Walmart’s lessons about basket behaviour can be used as a checkpoint. (WIRED) If your product content cannot support bundling, substitutions, and usage context, the assistant experience becomes fragile.
Retail experimentation in grocery, including Albertsons’ tool and Ocado’s agentic chatbot rollout, reinforces that this is not theory. (Grocery Dive) Retailers are moving. Brands that help them succeed will gain visibility.
Why this topic reignites momentum in careers as well as companies
Mid-level leaders often feel their work disappears into reports, meeting minutes, and decks. Agentic commerce creates a rare kind of cross-functional problem that senior leaders care about because it touches growth, margin, trust, and control all at once.
BCG’s projected shift in AI search traffic creates urgency. (BCG Global) McKinsey’s observation that AI already influences decisions in Europe confirms it is not “future talk.” (McKinsey & Company) Gartner’s prediction that most brands will adopt agentic AI interactions shows adoption will accelerate. (Gartner)
That combination creates an opportunity for visible leadership. The person who can translate this into a practical readiness plan, ship improvements, and create measurable trust signals becomes a go-to leader.
Company momentum returns when growth feels controllable again. Career momentum returns when your work becomes the work leaders rely on.
Closing
Agentic commerce will not replace retail overnight. It will not make websites irrelevant tomorrow. It will, however, change what it means to be “easy to choose,” because the first chooser may be an AI assistant long before a human ever sees your product.
The winning move is not panic. The winning move is readiness. Earn the shortlist by making product truth consistent, proof easy to retrieve, and decision logic simple to explain. That is how brands protect margin, and it is how mid-level leaders build the kind of influence that lasts.
Next Steps
Want a fast diagnostic of where your brand is most at risk in assistant-led discovery, and where you have an unfair advantage you can scale? Start with the QC2™ Evaluator, then use CATSIGHT™ to translate the results into the one change that will improve decision influence fastest.
