Introduction: From Search to Seamless Shopping
For most of the last two decades, online shopping began with a browser search and ended in a shopping cart. Consumers hopped between comparison sites, retailer pages and review blogs to evaluate products before finally checking out. That funnel is about to be disrupted. In late September 2025, OpenAI announced Instant Checkout for ChatGPT, an in‑conversation purchasing experience powered by the Agentic Commerce Protocol. Instead of acting as a helper that suggests product ideas, ChatGPT can now complete transactions on a user’s behalf, turning a chat window into a checkout lane. More than 700 million people turn to ChatGPT each week, and the service is the first large‑scale deployment of so‑called agentic commerce.
Agentic commerce describes a world where AI agents research, compare and even buy products for us. Visa, Mastercard, PayPal and Shopify are building payment rails and product feeds to support these autonomous shoppers. Perplexity’s Buy with Pro and Google’s upcoming “Buy for me” feature for AI Mode similarly allow single‑click purchases from within a chat interface. Analysts at Mirakl call this shift Commerce 3.0—a transformation where discovery, comparison and purchase collapse into one conversational flow, threatening to disintermediate merchants who rely on search and marketplace traffic. This blog explores how likely it is that AI platforms will enable purchases and what that means for the ecommerce and digital‑marketing landscapes.
ChatGPT’s Instant Checkout and the Agentic Commerce Protocol
OpenAI’s announcement marks the most concrete example of agentic commerce to date. Instant Checkout is available for U.S. ChatGPT Free, Plus and Pro users, starting with sellers on Etsy and expanding soon to over a million Shopify merchants such as Glossier, SKIMS and Vuori. Initially the system supports single‑item purchases; multi‑item carts and international expansion are on the roadmap. When a shopper asks for “best running shoes under $100,” ChatGPT returns a carousel of relevant products. If a product is eligible, a Buy button appears. The user taps it, confirms shipping and payment details, and the order is placed without leaving the chat. The merchant calculates tax, ChatGPT displays the total and collects a secure token, and the payment processor (usually Stripe) completes the charge.
Crucially, ChatGPT acts only as an intermediary. Orders, payments and fulfillment are handled by the merchant using their existing systems. The experience is free for users; merchants pay a small fee, but OpenAI stresses that product results remain unsponsored and ranked purely by relevance. When multiple merchants sell the same product, ChatGPT considers factors like availability, price, quality, primary seller status and whether Instant Checkout is enabled. Merchants can integrate with just a few lines of code thanks to the Agentic Commerce Protocol, an open standard that OpenAI co‑developed with Stripe and open‑sourced for developers. The protocol allows any AI agent to transact with any merchant or payment processor, meaning the technology could quickly expand beyond ChatGPT.
How It Compares to Other AI Shopping Assistants
ChatGPT isn’t alone. Perplexity debuted Buy with Pro in 2024; paying subscribers can purchase products recommended by its search engine and enjoy perks like free shipping. Google demonstrated its Agent Payments Protocol (AP2) in May 2025, which will let its Gemini‑powered AI Mode handle end‑to‑end checkout including price tracking and virtual try‑on. Visa’s Intelligent Commerce, Mastercard’s Agent Pay and PayPal’s Agent Toolkit provide secure payment frameworks for AI agents. European marketing consultancy Digital Loop notes that platforms like Check24, Amazon’s shopping assistant Rufus, and Perplexity + PayPal already offer in‑chat purchases.
Despite the hype, adoption remains limited. All current implementations restrict purchases to one item and to specific merchants. Google’s AP2 is still in demo phase and initially only supports U.S. merchants that accept Google Pay. Perplexity’s program covers a small number of U.S. merchants and is paywalled. OpenAI’s Instant Checkout works only with Etsy and soon Shopify sellers, though expansions are expected. Infrastructure constraints—fragmented APIs, varying checkout logic and lack of universal product taxonomy—make it difficult to support thousands of merchants.
Why Agentic Commerce Is Likely to Grow
Analysts generally agree that agentic commerce will grow rapidly once these limitations are resolved. Exploding Topics reports that 63 % of AI users already rely on ChatGPT for research. More than 700 million people use the platform weekly. When OpenAI added Instant Checkout, it instantly gave those users access to over five million Etsy stores. The same release promised that over a million Shopify merchants would be added soon, potentially in time for Black Friday. ChatGPT’s scale alone means adoption could happen quickly: what begins as a niche feature could become a default shopping channel for millions.
Other statistics underline the momentum. Mirakl cites a Gartner prediction that 40 % of enterprises will leverage AI agents by 2026 and notes that 60 % of shoppers already use AI in their shopping journey. Adobe’s research shows a 4,700 % year‑over‑year increase in traffic from AI search platforms. Meanwhile, Capgemini found that 71 % of ecommerce customers want AI capabilities integrated into their buyer journeys. In Digital Loop’s survey of executives, over 70 % of companies are already using or implementing AI agents. These figures suggest strong demand from both consumers and businesses.
Convenience will also accelerate adoption. The Xenoss report notes that users will likely trade small security concerns for convenience; only 8.5 % of surveyed users always trust AI Overviews, yet they continue to adopt AI tools because they save time. Agentic commerce reduces cognitive load and shopping time by matching products to highly specific queries, tracking prices and offering one‑click buying. It might even analyze personal taste from images or past conversations to propose the perfect gift.
Risks and Constraints
Technical and Regulatory Barriers
The infrastructure for agentic commerce is still immature. Xenoss points out that existing shopping workflows were designed for human users, not AI agents, leading to a lack of unified back‑end systems. Retailers use proprietary APIs with different formats and rate limits, there is no shared checkout logic, and product taxonomies are inconsistent. These fragmentation issues explain why OpenAI had to limit Instant Checkout to Etsy and why Google’s agentic checkout requires merchants to accept Google Pay. Xenoss estimates that it may take at least a year for the infrastructure to mature and support wide‑scale agentic commerce.
Data security is another concern. AI shopping agents handle sensitive payment information, so prompt‑injection attacks and model manipulation pose serious risks. Xenoss warns that unauthorized purchases, data breaches, compliance violations and account takeovers are possible if malicious actors exploit AI prompts. They emphasize that each action must require explicit user consent, payments should use encrypted tokens, and only minimal data should be shared. OpenAI’s Instant Checkout already follows these principles: users confirm each step and payment tokens are authorized only for specific amounts.
User Experience and Trust
Agentic commerce raises questions about trust and transparency. CMSWire points out that in a chat interface there are no “Ad” labels or visual cues to indicate sponsored placements, so customers might not know why certain products are recommended. Maintaining trust will require ongoing transparency about ranking algorithms and data use. Customers may also expect conversational support after a purchase—asking ChatGPT about order status or returns—but merchants still handle fulfilment and support, so there may be a mismatch between expectations and actual service. Companies will need to align their support channels with the conversational buying experience.
Market Implications and Competition
Agentic commerce threatens to disintermediate retailers. Mirakl warns that as AI agents take over discovery and decision‑making, traditional search and marketplace traffic could decline. Some retailers are already seeing a drop in organic traffic: nearly 10 % of businesses experienced declines due to generative AI’s erosion of search clicks. BCG projects that U.S. spending on AI search ads will reach $26 billion by 2029, about 14 % of total search ad spend. If AI platforms become gatekeepers, brands may lose customer data and the ability to cross‑sell; agents could break up multi‑item purchases across different retailers to optimize price.
There is also competition among platforms. Google, Visa, Perplexity, Amazon and others are racing to offer shopping agents, and each has its own payment rails and merchant eligibility requirements. As these platforms compete, they may introduce ads or sponsored listings. Search Engine Land expects that after this “organic” phase, paid placements will arrive. Merchants that integrate early will have a head start in organic rankings, but competition for visibility will intensify as AI ad products develop.
How Agentic Commerce Will Change Ecommerce
A New Shopping Funnel
Instant Checkout collapses discovery, comparison and purchase into a single conversation. OpenAI calls this a “chat‑to‑checkout” experience. Instead of searching multiple websites, users ask ChatGPT, view curated results and buy without leaving the chat. This drastically reduces friction, which could lead to higher conversion rates. Consumers will likely expect similar convenience on all platforms, pushing retailers to adopt AI‑driven assistants or risk losing sales.
Merchant Strategy: From Websites to Feeds
Agentic commerce shifts the emphasis from storefront design to data quality. To show up in AI recommendations, merchants need well‑structured product feeds with accurate titles, attributes, pricing, availability and reviews. Coalition Technologies advises merchants to optimize for AI SEO (also called Generative Engine Optimization or GEO) by providing rich product data, competitive pricing, high‑quality images and strong reviews. Search Engine Land explains that ChatGPT ranks products based on availability, price, quality and whether the merchant is the primary seller. Enabling Instant Checkout further improves visibility.
Merchants should prepare for multi‑item carts and cross‑sell capabilities. As ChatGPT adds support for carts, there will be opportunities to offer complementary items and increase average order value. However, there is also a risk that AI agents will pick individual items from different retailers to find the best deal. Retailers need to ensure that their own agents or product feeds can compete on price and service.
Building Owned Agents and Unified Data
Given the risk of disintermediation, many retailers are developing their own AI agents. Mirakl suggests brands can build bespoke shopping assistants that provide personalized experiences and collect customer data. For example, L’Oréal’s virtual beauty assistant recommends routines based on skin type. By owning the agentic interface, brands maintain the relationship and can offer unique services beyond commodity comparisons. Building such agents requires a unified data platform, scalable product catalog and robust operations to ensure accurate recommendations.
Security and Compliance as Differentiators
Retailers that prioritize security will build trust. Xenoss notes that comprehensive guardrails—explicit user consent, encrypted payments and minimal data sharing—are essential. Merchants can differentiate themselves by transparently communicating these practices and by offering easy returns and customer support. They should also prepare for regulation: as AI agents gain autonomy, governments may impose new rules on data sharing, liability and consumer rights.
How Agentic Commerce Will Change Digital Marketing
Rise of Generative Engine Optimization (GEO) and Agent‑Oriented Optimization (AOO)
Traditional SEO is built around ranking in search engines, but AI chatbots and agents have no “first page.” Generative Engine Optimization (GEO) focuses on making content discoverable and actionable for AI. Work & Co argues that GEO must be treated as a product‑design problem rather than a marketing tactic. Brands must design content and data so that AI systems can confidently interpret and act on it. This involves providing structured data, clear plain‑language content, and authoritative sources. GEO extends beyond customer acquisition into support, because AI agents will answer post‑purchase questions as well.
Digital Loop introduces Agent‑Oriented Optimization (AOO), a discipline that ensures product data is machine‑readable, up‑to‑date and richly annotated. An AOO manager enriches product titles and descriptions, supplies real‑time pricing and availability, defines return policies and warranty information, and provides identifiers like GTINs. These details enable AI agents to make informed decisions and, importantly, to trust a product enough to recommend it. AOO builds upon Answer‑Engine Optimization (AEO) and Generative AI Optimization (GAIO) but focuses on direct transactions.
Shifting Channels and Budgets
As AI‑driven interfaces gain traction, traffic from traditional search and marketplace channels may decline. Mirakl notes that AI search traffic grew by 4,700 % year‑over‑year, yet conversion rates are still catching up. At the same time, nearly 10 % of businesses have seen a drop in organic search traffic. Marketers will need to reallocate budgets toward AI platforms—eMarketer estimates that AI search ads could account for 14 % of total search ad spend in the U.S. by 2029. Search Engine Land predicts that once the organic phase ends, OpenAI will introduce paid placement options. Companies that master GEO and AOO early will gain an organic advantage before the paid layer arrives.
Importance of Trust and Brand Authority
In an AI‑led environment, brand recognition and trust become even more critical. Without the visual cues of traditional ads or product pages, consumers depend on the AI agent’s credibility. Work & Co emphasizes that brands must become trusted authorities in their domains. This involves producing clear, reliable content, earning positive reviews and engaging in thought leadership. Reviews play an outsized role: AI agents weigh quality and sentiment heavily when ranking products. Marketers must therefore foster genuine customer feedback and address negative reviews promptly.
Personalization and Data Privacy
ChatGPT’s Memory feature and custom instructions allow the model to remember user preferences and personalize recommendations. This creates opportunities for highly targeted marketing but also raises privacy concerns. Marketers will need to balance personalization with transparent data practices and user control. Regulations like the EU’s AI Act and evolving U.S. privacy laws could influence how AI platforms store and use consumer data.
Emergence of New Roles and Tools
As agentic commerce evolves, new roles like AI SEO specialists, AOO managers and agentic experience designers will emerge. Tools for prompt research, AI feed optimization and cross‑platform analytics will grow. Already, platforms like Semrush are launching AI visibility tracking tools that show which prompts surface your brand in ChatGPT, Google AI Mode and other engines. Digital Loop warns that brands that fail to adapt quickly could vanish from customers’ view. Investing early in agentic strategy will pay dividends as the field matures.
Conclusion: Preparing for the Agentic Commerce Era
Purchasing products through AI platforms like ChatGPT isn’t a speculative future—it’s already happening. OpenAI’s Instant Checkout enables U.S. ChatGPT users to buy from Etsy sellers today, with Shopify integration and multi‑item carts on the horizon. Other platforms such as Perplexity and Google are developing similar capabilities. Consumer adoption is likely to be rapid: billions of dollars in AI search traffic, high user trust in convenience, and a deluge of corporate investment suggest that agentic commerce will soon be mainstream.
Yet this new paradigm comes with challenges. Infrastructure fragmentation, security risks, limited merchant coverage and user‑trust concerns must be addressed before agentic commerce can scale. Retailers risk losing control over customer relationships and need new strategies to remain visible and relevant. Digital marketers must pivot from traditional SEO to Generative Engine Optimization and Agent‑Oriented Optimization, designing data for AI consumption and building brand authority in a world where conversations, not clicks, drive commerce.
The journey from chat to checkout is just beginning. Companies that invest now in structured product data, secure payment integrations, transparent customer experiences and AI‑optimized content will be ready for the next phase of ecommerce. Those who ignore the agentic shift risk being left out of the conversation entirely.