Agentic commerce

Agentic commerce is the practice of delegating online shopping tasks — product research, comparison, selection and sometimes checkout — to artificial intelligence agents acting on a shopper's behalf. As consumers increasingly ask assistants like ChatGPT, Gemini and Perplexity what to buy, product discovery is shifting from search engines and marketplace websites to zero-click AI-mediated channels, changing how brands reach customers.^[1]

Background

Agentic commerce emerged from the convergence of capable large language models, real-time product data feeds and one-click payment integrations. Through 2025 and 2026 the major platforms shipped the consumer-facing building blocks: Perplexity launched a free shopping experience with conversational product discovery, personalised product cards and PayPal-powered instant checkout; OpenAI introduced shopping research in ChatGPT, using a smaller GPT model trained with reinforcement learning to build comparative product guides; and Google rolled out agentic checkout across Search's AI Mode and Gemini — a "Buy for me" capability that lets agents complete purchases on merchant websites, live with selected retailers in the United States.^[1]

Analysts describe the shift in stages. Forrester's mid-2026 assessment found brands focusing on influencing AI agents through machine-readable content — building a "machine advantage" so agents favour their brand — rather than waiting for fully autonomous end-to-end journeys.^[2] Practitioner research similarly argues that early deployments are purpose-built agents (reorder automation, bundle builders, shopping assistants) instead of a single monolithic agent orchestrating the entire customer journey, with agent-to-agent protocols still emerging.^[1]

Key concepts

• Zero-click commerce — the shopper never visits a merchant site; the agent handles discovery and may complete checkout, so visibility depends on the AI platform's retrieval rather than ranking in search results.^[1]
• Answer engine optimisation (AEO) — the successor discipline to SEO: structured product data, enriched metadata and clean catalogues determine whether an agent can understand and recommend a product.^[1]
• Agent checkout — payment authorisation delegated to the agent, requiring new authentication, fraud and liability frameworks; a majority of financial institutions expect fraud pressure to rise as purchasing agents proliferate.^[4]
• Merchant dependency — as AI platforms control the discovery layer, organic search traffic to retailers declines and brands grow dependent on the platforms that own the consumer interface and payment rails.^[3]

Applications and impact

Early use cases concentrate on discovery and conversion: comparative product research, price monitoring, replenishment, and B2B procurement, where one in five sellers may soon need to respond to AI buyer agents with counteroffers from their own seller-side agents.^[1] For retailers, the implications run from catalogue engineering to loyalty — brands that make their product data legible to agents gain visibility, while those that do not risk disappearing from consideration. The model also redistributes economics: the platforms owning the interface and checkout capture new leverage over merchants, and consumer adoption is already broad — OECD data shows more than one-third of individuals across member countries used generative AI tools in 2025.^[5]

>Key Takeaways

• Agentic commerce moves shopping discovery from search and marketplaces to AI assistants that research and buy on the user's behalf.
• Perplexity, OpenAI and Google shipped shopping and agent-checkout features through 2025–2026; brands respond with answer engine optimisation.
• Merchants face declining organic search traffic and new dependency on the AI platforms that own the consumer interface.

See Also

• Agentic AI
• AI Agents
• Agent2Agent Protocol
• AI in Malaysian Retail
• AI in Malaysian SMEs