At this year’s Consumer Goods Forum Global Summit in Vienna, retail and consumer goods leaders are moving beyond the question of whether AI may change the consumer journey. The focus now is speed, how quickly companies can adapt their data, supply chains, and commercial relationships to stay visible as algorithms begin shaping more purchase decisions.
By Evan Sheehan, Deloitte Global Retail, Wholesale & Distribution and Consumer Products Leader and Ed Johnson, Principal, Deloitte Consulting LLP
Retailers and CPG brands planning for a return to stable inflation and predictable supply chains may be optimizing for a world that no longer exists. This is not navigating temporary disruption, but navigating an era of compounding instability.
For decades, retail growth was built around a familiar playbook: win the right shelf space, invest in distinctive packaging, and use promotions to drive conversion and shape consumer choice at the point of decision. That playbook was built for human behavior, and it assumed shoppers could be influenced by endcaps, price cues, loyalty offers, packaging claims and familiar brand signals. Those levers still matter, but on their own, they may no longer be enough.
The next structural shift in commerce is already taking shape. In more shopping journeys, the “shopper” may not be a person walking an aisle or scrolling a product page — it may be an AI agent researching options, comparing value, building a basket, or completing a purchase on the consumer’s behalf. This is the rise of the invisible shelf, a world where visibility is determined not only by physical placement, search ranking or paid media, but by whether a product can be understood, compared, trusted, and selected by algorithms.
The universal squeeze
This shift is arriving at a difficult moment. Deloitte’s ConsumerSignals data shows 73% of consumers globally expect grocery and fuel prices to keep rising, reflecting a consumer base that isn’t confident but is actively coping with a challenging environment. Value-seeking behavior has also moved upmarket. A significant share of higher-income households are now engaging in multiple value-seeking behaviors alongside lower-income households.
At the same time, essential spending remains elevated while discretionary intent stays under pressure, and consumers are scrutinizing price, quality, availability, and benefits more carefully before they buy. That is putting pressure on the old growth model. For years, many brands could grow simply by being broadly acceptable, not the cheapest, not the most differentiated, but rather familiar, available and frequently promoted. That middle ground is becoming harder to defend, as consumers at one end trade down in search of affordability while those at the other maybe remain willing to pay more once the value equation is clear.
According to Deloitte’s Value-Seeking Consumer research, depending on the category, price may explain 40% to 60% of perceived value, leaving a significant role for attributes such as taste, performance, convenience, claims, provenance, and trust. It’s this undifferentiated middle that has become the real danger zone. If a product cannot clearly prove why it deserves to be chosen, it may risk being squeezed out entirely.
When AI becomes the point of decision
Agentic commerce could intensify that squeeze further. AI agents may not respond to the same signals that influence a person browsing a store. A product positioned at eye level may not trigger an impulse purchase, fine packaging may not earn a second glance, and a promotion may not register as good value unless the underlying data actually supports it.
Instead, agents optimize. They compare based on metrics like mathematical value, real-time fulfillment, dietary needs, household preferences, ratings, product attributes and other data points, and they are likely to favor whichever products can prove their value clearly and consistently. That does not mean brands become irrelevant. In fact, it may make trusted brands even more important. But it does change how brand value should be expressed. A product’s benefits can no longer live only in a campaign, a trade story or a packaging claim; they need to show up in structured, accurate and consistently available data across the commerce ecosystem.
The potential risk is the rise of “ghost brands,” products that still exist on the physical shelf but may disappear from algorithmic consideration because their value, availability or attributes cannot be read and trusted by the systems making recommendations.
From joint business planning to agentic orchestration
This is where the next phase of retailer and supplier collaboration must begin. Traditional joint business planning has often been built around annual cycles, commercial negotiations, and promotional calendars. Those activities should remain important, but they were designed for a slower, more predictable model of commerce, and the invisible shelf demands a more dynamic framework in its place.
The starting point is master data. Product descriptions, ingredients, allergens, pack sizes, claims, certifications, pricing, availability, and fulfillment options need to be accurate and structured in ways that AI systems can interpret. Poor data has created friction, but in an agentic world, it could mean outright exclusion from consideration.
Availability matters just as much, because supply chain agility can be difficult to fake convincingly to an algorithm. If an AI agent is optimizing for the best outcome on behalf of a household, it may prioritize products that can be delivered or collected dependably. A brand with high awareness can still lose out if the system can’t trust its availability and recommends an alternative instead.
Retailers and suppliers may need sharper value architecture, with each product assigned a clearer role. Is it designed to win on affordability, or does it command a premium because of superior quality or a specific consumer need? The least defensible position remains the unclear middle, where neither the algorithm nor the consumer sees a strong reason to choose.
Finally, collaboration needs to shift from periodic planning to a continuous rhythm. Demand signals, pricing, inventory, and consumer preferences now may move too quickly for legacy processes to keep pace. Retailers and CPG brands may need more frequent, more precise and more interoperable data exchange, all supported by appropriate governance.
Winning the algorithm to win the customer
None of this means the human element disappears. Consumers everywhere still care about taste, trust, quality, service, and experience, and stores still matter, especially when they create discovery, immediacy, and reassurance. But the path to the consumer is changing, and in many categories, the first filter may increasingly be algorithmic. If a product does not pass that filter, it may never reach the human moment of consideration.
For retail leaders, the shelf is no longer only a place; it is a system of data, decisions, and fulfillment promises. For CPG leaders, the next advantage may not come from producing more content or running more promotions, but from making each product’s value proposition clear, machine-readable and operationally reliable.
The companies that adapt fastest will ask a more fundamental question, when an AI agent is acting in the consumer’s interest, why should it choose our product or brand? The goal is still to prove value to the human, but increasingly, the way to get there is to first win the math with the bot.
