The fourth quarter of 2025 saw several high-profile partnerships between major large language model (LLM) companies like OpenAI and Perplexity and retailers to allow consumers to access retailers’ products in LLM chats. E-commerce is shifting from search and scroll to conversation and conversion. What was once experimental is now operational, and retailers must reconceive how they win attention, influence purchase decisions, and fulfill orders in an AI-mediated world.

A tidal shift in online shopper engagement

For decades, digital commerce centered on search engine optimization (SEO), paid search, and marketplace rankings. Retailers optimized site search, user experience, and navigation to keep customers on property, while retail media networks let brands buy visibility. Now, AI agents are acting as intermediaries between shoppers and retailers, routing discovery through conversational interfaces rather than search results pages. 

That change upends the rules of visibility. Influence will be earned through relevance in AI recommendations rather than placement on a results page. To keep products discoverable, retailers must make product data machine-readable, surface real-time inventory to AI engines, and cultivate brand preference so shoppers mention products by name when interacting with assistants. The battleground is shifting from SEO to what we might call GEO — generative engine optimization.

Conversational recommendations reshape promotions

Early partnerships between agents and mass retailers — examples include Instacart, Walmart, and Shipt — target longtail assortments where browsing is often overwhelming. In these contexts, AI simplifies discovery and can surface items with a single natural language prompt. 

Promotions that once relied on banners, cross-sell widgets, or limited time offers must evolve. In a dialogue, an AI can proactively suggest a complementary product or an entire meal plan based on one item query. Promotions then become embedded within the conversational flow rather than appended to a page.

Retailers and brands must learn to craft contextual offers that feel native to the conversation and to feed AI systems with scenarios and associations that expand baskets organically.  Today’s interactions lack the power of suggestive merchandising.

The trade spend model is due for AI-driven reinvention

The underlying commercial fabric of retail — trade spend and shopper marketing funds — will be disrupted. Historically, these funds have flowed to in-store displays, retailer media, and paid placements; with the chat providers acting as AI intermediaries, influence is initially earned through product quality, relevance, and positioning. That dynamic is unlikely to persist indefinitely: Chat providers will begin to compete for a share of trade dollars, and new commercial programs will emerge to optimize for AI discovery. 

Brands and merchants should start designing programs that collaborate with retailers and platform providers to present products in ways that win under conversational rules. Joint business planning among suppliers, retailers, and AI intermediaries will become a core competency rather than a nice to have.

Conversational engagement reduces costs and fuels loyalty reinvestment

Conversational assistants compress discovery, decision, and checkout into a few conversational turns, increasing expectations for seamless, fast fulfillment. This compression will likely produce more frequent, smaller orders with less predictable timing, a strain on traditional fulfillment models. 

The opportunity is to use conversational engagement to smooth demand. Retailers can unlock supply chain efficiencies by moving shoppers from last minute, “What’s for dinner tonight?” type decisions to efficient weekly planning through conversation. Services such as meal planning, routine restocks, and bundled baskets can turn conversational moments into predictable demand patterns, reduce cross shopping, and lower operating costs — enabling reinvestment in pricing, experience, and loyalty.

A generational change in shopping behavior

Just as Gen Z embraced TikTok as a discovery channel, the cohort whose formative digital experiences began with ChatGPT will expect commerce to be woven into conversations. This generation will view AI assistants not only as guides but as primary purchase channels, creating new buying occasions — shopping while planning a week of meals, for example, or restocking during casual exchanges. Retailers that rapidly experiment with conversational selling models and design experiences that feel native to AI interactions will be best positioned to capture long-term loyalty and market share.

Hybrid shopping interfaces — combining chat and visual basket management

While chat is an intuitive starting point, it won’t be the whole story. For large grocery orders or complex purchases, shoppers will need hybrid interfaces that seamlessly toggle between dialogue and visual curation.

Design systems that enable fluid movement between conversational prompts and visual basket management will better mirror how people shop today and will be essential for high-velocity, high-value categories.

In practice, this looks like a live chat that produces a compact, real-time basket preview users can tap to expand into a visual canvas with product photos, drag and drop reordering, bulk edit controls and contextual substitution suggestions; chat remains available in a sidebar so shoppers can ask targeted adjustments and immediately see cost, nutrition and availability deltas.

Exhibit: Operational implications for retailers

Source: Oliver Wyman analysis

Conversational commerce — the imperative to act with agentic AI

Agentic AI — systems that proactively engage, recommend, and transact — is no longer hypothetical. Sitting on the sideline’s risks loss of visibility at the moment consumers decide. Retailers should move from experimentation to deliberate capability building: Enrich data, prototype hybrid interfaces, rethink promotions for dialogue, and redesign fulfillment for conversational demand. 

The future of retail will be written inside model mediated conversations. Retailers that show up with the right product data, promotional playbooks, and fulfillment muscle will turn conversation into commerce; those that may find consumers simply talking past them.