Before you dive into trying to optimize your eCommerce store for AI engines, you must understand how these models work. Let’s dive into what agentic commerce looks like across different AI engines:

Open AI

The first product experience built on the ACP is Instant Checkout in ChatGPT, designed to manage the entire buying journey without requiring the customer to leave the conversation.

Step 1: User query and intent recognition

User submits a request with clear commercial intent to ChatGPT, often through a Custom GPT or the integrated Operator.

Step 2: LLM tool selection and product feed integration

The LLM detects the need for commerce data and maps the request to Commerce APIs using function calling. Brands must supply structured product data (name, description, price, availability) as the product feed.

Step 3: Product presentation and selection

The agent identifies relevant, in-stock items and displays personalized recommendations directly within the chat interface.

Step 4: One-click transaction initiation via ACP

When the user chooses a product, the Agentic Commerce Protocol launches a native checkout session inside the AI environment, removing friction from traditional checkouts.

Step 5: Secure payment processing

Payments are processed natively via Stripe infrastructure, supporting saved payment methods through Stripe Link.

Step 6: Merchant authorization and fulfillment

Order details are securely transmitted to the merchant’s systems. The merchant authorizes the order and handles fulfillment and delivery.

OpenAI’s decision to launch the ACP immediately and integrate it into ChatGPT represents a strategy of establishing a de facto standard through rapid product adoption rather than waiting for industry-wide consensus. This approach prioritizes creating a fast, direct path to the consumer, transforming the chat platform into a direct sales channel with minimal friction.

Google AI shopping and the autonomous procurement model

Google’s approach to agentic commerce focuses on building an open, foundational utility standard that leverages its dominant search position and vast data catalog to automate procurement on behalf of the user.

Beyond basic listing retrieval, Google’s AI Shopping incorporates advanced agentic capabilities to solve key consumer shopping problems:

• Virtual try-on (VTO): This technology addresses the critical challenge of knowing how clothing will fit. Shoppers can tap the "try it on" icon on product listings, upload a full-length photo, and the AI will apply the apparel (shirts, dresses, pants) to the user's photo, preserving subtle details of fit and drape.

• Shoppable visual inspiration: The AI Mode allows users to explore visual concepts, such as designing a room layout or creating an outfit, and matches those visions with specific, shoppable products.

  
  

Perplexity and the transparent answer-engine model

Perplexity focuses on deep conversational understanding and unbiased ranking, positioning itself as a highly trustworthy AI-powered answer engine for commerce.

Step 1: Capturing complex, multi-attribute intent

The agent translates messy queries like "quiet blender under $150, fits small kitchens" into precise vector attributes such as price, decibel rating, capacity, and footprint for accurate, contextual search.

Step 2: Building the answer (real-time data synthesis)

The engine calls tools to gather live data from merchant feeds, price APIs, community chatter, and reviews, synthesizing specs and public sentiment.

Step 3: Reasoning and ranking (unbiased logic)

GPT-4 logic evaluates trade offs and ranks products by relevance and data quality. No paid placements influence results, so brands with accurate, trustworthy data rank higher.

Step 4: Checkout and fulfillment

• Buy with pro native checkout: One-click native checkout for eligible pro users in the U.S., autofilling transaction details via merchant APIs.

• Merchant redirection: If native checkout is unavailable, the user is sent to the merchant site to complete the purchase.

The Perplexity Shop workflow emphasizes analytical rigor and real-time data synthesis, often using GPT-4 logic for informed reasoning.

The consequence of Perplexity's unbiased ranking is a fundamental shift in e-commerce optimisation. Product visibility now depends heavily on Product Data Quality Optimization (PDQO). Retailers must ensure their product detail pages (PDPs) and data feeds are comprehensive and strategically structured to satisfy the AI’s rigorous reasoning criteria, rather than relying solely on traditional SEO or paid media tactics.

Our take: Agentic commerce isn’t about chasing AI engines in hopes of getting mentioned. It’s about making AI chase your brand to rank. Fix your foundations by cleaning up your product data, uploading relevant product content and maintaining reviews.