Mars Redesigns Impulse Strategy as AI Agents Take Over Grocery Runs
The Development
Mars is hedging against an agentic retail future in which AI assistants take over a larger share of the weekly grocery run. Neil Reynolds, the company's global chief customer officer, told Digiday he does not expect AI to eliminate the in-store checkout lane, but does expect agentic shopping assistants to absorb more routine basket-building. The stance rests on Mars' recent future of shopping and retail report, produced with Kantar and Synthesis. The company owns Dove chocolate, M&Ms and Pop-Tarts, categories where a large share of volume comes from unplanned purchases. Reynolds described Mars' effort to sell products to AI systems, the value of visibility inside those systems, and what happens to impulse purchasing when software, not a person, fills the cart.
Our Take
Mars is right to separate the weekly shop from the checkout lane, and the split is the real insight. Routine replenishment is exactly what agents will own first, which means the planned basket becomes a data-and-eligibility contest while impulse survives only where a human is physically present. Most CPG peers are still debating whether agentic commerce is real. Mars is already allocating research budget to it, and that is the signal. The strategic risk is not losing shelf space; it is never being surfaced by the agent at all. Brands that treat AI visibility as a future SEO-style project will find the default choices already set.
What Changed
Shopping agents can now execute the recurring grocery basket autonomously, selecting products against stored preferences and constraints. That moves the purchase decision from a human browsing shelves to a model ranking options, so brand selection becomes a machine-readable optimization problem rather than a persuasion problem.
Marketing Impact
Shopper marketing and trade teams lose a lever: displays, endcaps and promotional pricing aimed at humans do not move an agent. Ecommerce, product data and retail media teams gain the central role, because structured attributes, reviews and price signals now determine whether a brand enters the agent's consideration set.
Competitive Implication
Large CPG groups with first-party shopper research and retailer data partnerships gain structural advantage, because they can test what agents actually select. Challenger brands relying on in-aisle discovery and promotional visibility are vulnerable: agents optimizing on price and past purchase history entrench incumbents and starve unplanned trial.
Strategic Outlook
Expect other major CPG companies to publish similar agentic-shopping research within two quarters, followed by retailer-agent partnerships that trade access for data. By 2027, checkout-lane and last-mile impulse tactics become the protected remainder of the category, while core replenishment shifts to agent-optimized listings.
The Exploit
Action Item
Heads of ecommerce at CPG brands should audit product attribute, ingredient and availability data for agent readability before Q1 2027 planning closes, then run controlled prompts across major shopping assistants to measure how often their SKUs are selected against rivals.