Earlier this month, OpenAI scaled back plans to enable in-chat purchasing through its Instant Checkout feature. For retail leaders who feared this next big wave of disruption (or didn’t believe in it in the first place), this might seem like welcome news.

But it’s far too early to call this the end of the agentic commerce era. Other announcements this month from Sephora and Carrefour demonstrate that momentum is building, even if execution is uneven. Our recent research revealed that 38% of shoppers already trust AI agents to manage routine purchases and 55% would be willing to let agents handle reorders in the next three years.

So what happens next? The answer might be found by looking to the past.

I expect agentic commerce to follow a similar path to marketplaces. When platforms like DoorDash and Instacart first emerged, many retailers were hesitant to participate, wary of ceding control over customer relationships and data. However, once a critical mass of consumers was shopping on that channel, retailers realized that by not being present, they were potentially giving up a sizeable slice of the market.

As agentic commerce advances toward scale, retail leaders should use this time wisely, taking steps to address the shifting requirements in discovery, pricing, cost management, and loyalty that agentic AI will bring.

Here I offer 4 ways that retail will change in the agentic age how the industry should respond:

1. Product discovery shifts from grand-first to need-first

The most recent holiday shopping season data reveals an incredible surge in AI-driven traffic, with visits rising 693% year over year and revenue per visit from AI also growing by 254%.

In our most recent consumer trends survey, What matters to today’s consumer, we see why:  69% of shoppers say they trust a digital assistant to suggest new products or deals if the assistant explains its recommendations.

Shoppers clearly value the convenience of AI tools – and show strong trust in the research capabilities and recommendations they provide. But as retail AI agents take control of consumer product discovery, traditional SEO is being replaced by GEO, which will fundamentally re-define how brands get found, chosen, and bought.

The main issue brands face with GEO is objectivity. AI LLMs are more rational actors than human consumers. When shopping tasks are completed by AI tools, brand preference gives way to product fit and value for money. Having the best marketing alone will no longer define the market leaders. Instead, having the best product for the price and other “fact-based” factors is what will drive AI recommendations. 

We already see this behavior among savvy consumers, especially in categories like cosmetics, where TikTok influencers build massive followings by identifying affordable “dupes” for high-end products, the likes of which often sell out after just a single high-profile post. Agentic AI shopping assistants could potentially bring this dynamic to every category, helping consumers surface and purchase the best option regardless of past preference.

2. Pricing evolves from static lists to real-time AI negotiation

Consumer shopping assistants are already changing what shoppers will buy, but the next waves of agentic commerce could also see changes in how they buy.

While human shoppers value convenience and therefore limit their number of transactions, AI shopping assistants have no such constraints. They can present an endless combination of orders to optimize for cost, availability, fulfillment speed, quality, or freshness.

Think of it like the responses you get when using a generative AI tool like ChatGPT to write an email. Rather than offering a single response, the assistant often presents several options – each slightly adapted for tone or style – and recommends the best one. Agentic commerce may operate in a similar way, offering recommendations for how to split the order to optimize against multiple factors.

For some retailers, this could spell big trouble. For example, in a low operating margin sub-sector like grocery, which typically operates on 3-5% margins, traditional winners have built dominance through the economies of scale. Size of basket matters in maximizing this efficiency, especially when fulfilling eCommerce orders. An agentic-enabled consumer may pose a real challenge to these grocers, as AI tools could drive smaller basket sizes, each optimized by different factors like price and delivery speed.

In response, grocers will need to build the capabilities that will allow them to shift from static pricing to real-time negotiation to win the basket. One tactic we see regularly today that aims to win the whole basket is a fairly blunt instrument like monthly spend challenges that offer a cash back award once the goal is met. Over time, we are likely to see more dynamic AI pricing tools on the merchant side to “bid” on the basket in real time at the point of purchase, optimizing pricing and offers to win baskets that meet margin targets or incentivize shoppers to increase their spend.

3. Loyalty moves from emotional engagement to rational performance

In an agentic commerce landscape, the traditional emotional loyalty loop is disrupted as AI agents prioritize cold, hard data over brand affinity. To remain relevant, retailers must pivot from human-centric visual persuasion to machine-readable value.

For example, retailers must ensure loyalty benefits, such as member-only pricing, real-time availability, and “agent-exclusive” shipping windows, are recognized by model context protocols (MCPs), such as Google’s Universal Commerce Protocol, and can be factored into a rational utility calculation.

Since agentic commerce turns shopping into a “needs-first” mission, loyalty must become a utility rather than a hobby.

To do this, retailers must embrace three key ideas:

The “trust premium” reward

Reward “consistency” by offering agents guaranteed fulfillment slots or “green light” shipping (instant dispatch with no fraud-check delay) for recognized, high-frequency identities.

Non-transactional access

In a world of automated buying, human connection happens outside the purchase. Offer VIP access to physical experiences, community events, or early product testing — benefits an agent can’t “consume,” but the human owner values.

Identity-as-a-key

Allow the consumer’s agent to act as a universal wallet, where the loyalty status isn’t just a discount but a verified credential that unlocks faster service or higher-tier customer support across a partner ecosystem.

As a result of this shift, creative differentiation can also move beyond points to functional rewards, such as automated “gap-filling” (where a retailer’s agent might suggest a $2 add-on to hit a $5 discount threshold) or predictive post-purchase excellence (like autonomous proactive returns or “concierge” maintenance alerts), essentially training the consumer’s agent to favor the retailer based on service offering and seamless operational integration.

4. Cost optimization accelerates through autonomous and augmented AI

Inflation rates across North America and Europe may have eased slightly in recent months, but consumers are still feeling the pinch, especially as gas prices soar. In a February 2026 survey by Talker Research, 87% of consumers agreed that the U.S. is in an affordability crisis and 50% reported they struggle to afford necessities, such as groceries.

As shoppers use AI agents to compare and negotiate prices in real time, passing on cost increases becomes even harder for retailers, adding pressure to margins. Similarly, ‘shrinkflation’ tactics are more easily spotted by a ‘fact-based’ agentic shopper and thus harder to deploy.

These dynamics accelerate the need for retailers to find new ways to drive significant cost takeout and operational efficiency. AI offers a powerful opportunity to do both. However, to unlock its potential, retailers must go beyond surface-level tweaks and fundamentally rethink their operating models and end-to-end value chains. This requires a zero-based process redesign, applying AI where it can have the greatest transformational impact rather than just patching existing workflows.

Achieving maximum impact also demands a blended approach that integrates two distinct types of AI. First, autonomous or agentic workflows enable functions that the human workforce simply does not have the capacity to achieve today. For example, agentic AI could enable intelligent real-time supply chain synchronization by reacting to demand sensing at a highly granular level, allowing for just-in-time co-ordination between suppliers, warehouses, transportation, and stores that maximizes on-shelf availability while minimizing excess inventory. Similarly, in-store autonomous systems could monitor and trigger proactive maintenance alerts for equipment such as HVAC and refrigeration, preventing costly breakdowns and food spoilage/shrink.

The second opportunity lies in the augmented worker, or equipping employees with a personal AI companion. This persistent, always-on agent augments an individual knowledge worker’s daily effectiveness, creating a win-win in enhancing both productivity and the quality of output. By empowering employees with AI co-pilots, retailers can reduce initial upskilling/training time, increase accuracy and quality of work outputs, and reduce overall time-to-completion for tasks, dramatically increasing the effective capacity of their existing workforce.


Preparing for retail’s agentic era: What leaders should do now

Agentic commerce is coming. How soon remains to be seen, but it will most certainly be part of the industry’s future. What retail leaders should do now is use their time wisely, adapting discovery, pricing, operations and loyalty strategies to meet the needs of all shoppers—human, agentic and whatever version may come next.

Discover how leading retailers are using AI to optimize pricing, operations, and loyalty in an agent-driven world. Connect with our retail AI experts to explore practical next steps.