Agentic AI and the New Consumer Frontier

I remember when search was king. If you wanted to understand consumer behavior, you looked at queries: “best running shoes,” “affordable skincare,” “holiday rentals in Italy.” Optimizing for keywords was both science and art, the backbone of performance marketing.

But as I scroll through today’s data, I see something inherently different. Over 60% of consumers now use AI assistants like ChatGPT for shopping queries, and nearly the same share report they no longer rely on traditional search engines at all. Within months, a consumer behavior pattern that took decades to cement is being rewritten.

AI Shopping Assistant usage by demographics reveals Gen Z has a higher propensity to leverage new technologies in their shopping journey:

The consequences ripple across the retail ecosystem:

• Retail media networks at risk. Built on keyword targeting, they may no longer map to how consumers discover products. If my AI assistant finds me “a waterproof hiking boot under £150 that works for Scottish winters,” there is no keyword bid behind that.
• Marketplaces redefined. With initiatives like ChatGPT’s checkout, answer engines are experimenting with becoming full marketplaces. If consumers complete the transaction within the AI environment, who owns the data? Who controls loyalty? Who captures the margin?
• The end of endless scroll. Gen Z consumers often bypass search engines altogether, starting their discovery on platforms like TikTok or Instagram. Their patience for sifting through hundreds of irrelevant results is near zero. They expect precision, relevance, and fun.

For me, this echoes lessons from when I started exploring marketplaces years ago. Then, the question was: how do we reach consumers in environments they already trust? Today, the question is sharper: how do we prepare for a world where traffic comes less from humans and more from AI agents acting on their behalf?

The answer lies in agility. Brands must think beyond keywords. Product content must be AI-readable and conversational, with lifestyle imagery that shows who the product is for and how it is used. D2C sites, once neglected, need to be optimized for crawling by answer engines. Freshness and taxonomy matter more than ever.

Commerce is no longer about being seen in search results. It is about being understood by machines.

If the collapse of search represents the micro-dynamics of shopping, Deloitte’s six forces of the consumer industry reveal the macro context. These forces: demographics, cultural shifts, technology, industry disruption, climate, and politics remind us that commerce is not just about technology but about the world we live in.

One force I find particularly striking is demographic wealth concentration. Baby boomers, once seen as a declining cohort, are now the largest contributors to consumer spending. Since 1990, wages for providers of labor have grown a modest 23%, but the wealth of asset-owning households (predominantly older) has increased by more than 1,300%. Add to this the projection that by 2030, women will control 62% of US wealth, and you see a radical shift in who holds purchasing power.

This is not just a story of income brackets; it is a story of values, priorities, and aspirations. Older consumers with asset-driven wealth may prioritize health, security, and legacy. Younger, asset-light generations face constraints but also value sustainability, secondhand markets, and shared ownership models. Mass strategies built on “the middle” no longer work in this fractured landscape.

Here, AI becomes transformative. For decades, personalization was constrained by economics: it was simply too expensive to localize assortments, to predictively stock stores, or to create truly dynamic pricing. But as Deloitte’s futurist Casey Lobaugh notes, AI dissolves those constraints.

The shift is from insights (human hypotheses about what consumers want) to predictive enterprise models (compute-led signals that surface patterns no human would see). I was struck by one example: ownership of a cat turned out to be highly predictive of new consumer acquisition for a retailer. It seems illogical, even comical, but predictive models don’t require human logic, they require outcomes.

Reflecting on this, I see both risk and opportunity. Businesses built for scale now face fragmentation. But fragmentation is not chaos if we have the tools to predict and respond at speed. The predictive enterprise is not about guessing the future; it is about building the agility to meet it.

Numbers and macro trends tell one story, but psychology reveals another: the inner calculus of trust.

Professor Luca Cian of the University of Virginia calls this the “Word of Machine” effect. His research shows that consumers are comfortable letting AI guide utilitarian purchases like insurance, household goods, financial services, but resist AI for hedonic purchases, such as luxury, art, or experiences. Consumers still believe humans better understand emotions and taste.

This duality fascinates me. On one hand, AI is freeing consumers from the paradox of choice. No more scrolling through 1,000 toothbrushes or 500 kinds of bottled water. The cognitive relief is real, and adoption rates prove it. On the other hand, when the purchase is about identity, pleasure, or creativity, consumers hesitate. They fear machines lack “soul.”

In fact, statistics show that consumers are happy to use AI shopping assistants to answer product queries:

Yet framing changes everything. Netflix recommendations are powered by algorithms, but consumers accept them as “people like you also watched.” Loyalty plays a similar role; it shortcuts decision-making, bypassing the need for awareness or consideration. If I already trust a brand, my AI agent doesn’t need to persuade me.

What intrigues me most is the emotional bond forming with AI itself. When OpenAI briefly retired GPT-4, many users described feeling bereaved, as if a trusted companion had been taken away. Resistance to AI’s role in hedonic decisions may soften as people grow accustomed to treating AI agents as companions rather than tools.

For leaders, this means that the psychology of adoption cannot be ignored. We are not just optimizing for efficiency. We are designing experiences that blend trust, loyalty, and human touch with algorithmic power.

So where does this leave leaders trying to navigate the storm? From my perspective, several imperatives stand out:

• Design for AI discovery. Ensure product data, imagery, and taxonomy are optimized for conversational AI. If an agent cannot “read” your product, it will not recommend it.
• Experiment with emerging marketplaces. Test partnerships with answer engines and be ready for a world where transactions no longer originate on your own site.
• Rethink segmentation. Move beyond demographics toward predictive signals. Who your consumer is may matter less than what patterns forecast their next decision.
• Balance efficiency with values. Today’s consumers equate “value” with alignment to sustainability, inclusivity, and convenience. Price is only part of the equation.

Humanize AI. Reframe agents as companions and co-creators, not competitors. Brands that blend human empathy with algorithmic precision will build deeper trust.