AI agents are beginning to transform how we shop, how brands sell, and how the entire stack of online commerce operates. Whether it is Perplexity shopping or OpenAI’s partnership with Shopify that will let you check out straight from ChatGPT, this change in consumer behaviour is happening quicker than people realise.
We are entering the Agentic Commerce Era: a world where AI doesn’t just recommend, but perceives, interprets, decides, and acts on your behalf. This shift has already begun with Google’s AI overviews, which have dramatically reduced organic traffic to publishers and ecommerce sites. And soon, we’re likely to see more AI Agents shopping on our behalf. It wouldn’t surprise me if in 10 years time, there will be more AI agents buying from ecommerce sites than actual users browsing and checking out than humans.
What Are AI Agents?
Most AI tools today are like calculators: powerful but reliant on human input. AI agents take this further. They work independently toward a goal, adapting to conditions and learning from experience.
OpenAI’s Operator, released in 2025, is a good example. It can browse online stores, compare products, and complete a purchase without step-by-step instructions. OpenAI has also partnered with Shopify to integrate direct checkout, meaning an agent can not only find what you want but actually buy it for you within the same workflow. This is the shift from “you shop” to “it shops for you.” I wrote about the basics of AI agents in a previous post, in case you want to go deeper.
AI Agents Already in Action
The impact of AI agents is already being felt across various aspects of ecommerce and enterprise workflows:
- Workflow automation: Harvey, the legal AI startup, has expanded from research to predictive analysis, drafting, and even negotiation. In retail, similar agent-driven automation is emerging across fulfillment, procurement, and logistics.
- Customer service: Companies like Sierra and Shopify Magic provide round-the-clock support, respond to messages, and trigger post-purchase communications. Shopify Inbox now handles customer conversations across channels. Businesses report that AI-powered service resolves the majority of issues, reduces costs, and drives higher sales.
- Personalized shopping: Agents learn preferences in detail, from clothing sizes and color choices to browsing history and body type. Generative AI can even adjust tone, images, or page layouts in real time, making every interaction feel personal and relevant.
- Marketing and content: Agents generate and continuously update product descriptions, optimise ads using live customer reviews, and manage campaigns based on real-time demand. This slashes the cost and time of traditional design and marketing workflows. There are many startups operating in this space such as Alison AI for creative, Munch Studio for automated social media content creation, etc (full disclosure: I’m an investor and board member at Munch with Remagine Ventures).
- Inventory and pricing: AI forecasts demand, prevents stockouts, and makes dynamic pricing adjustments based on market trends and competitor behavior, protecting margins while staying competitive.
- Conversational commerce: Voice and chat assistants are enabling true “shopping by conversation.” A prompt like “Find me new sneakers, same style as last time, half a size bigger, under $120” can be executed directly by the agent.
Even in hardware, startups such as Figure AI (humanoid robots) and Skild AI (robotics foundation models) are exploring how agentic systems could extend into the physical world.
The Future of Shopping
Looking forward, agents are poised to reshape not just the mechanics of shopping but the structure of commerce itself. A16Z recently published a new thesis on AI x Commerce describing where AI is being implemented in the customer journey, from AI agents used to find, research and compare deals for consumers.
- Agents as the new interface: Shopping will move away from search engines and browser tabs toward context-aware assistants. Buying will increasingly happen in the background, triggered by intent rather than explicit search. I touch on this in my post on AX vs. UX and how design is changing to accommodate this.
- Delegated purchasing: Agents will handle entire shopping missions, from research and comparison to ordering and returns. For routine essentials, they will monitor prices and reorder automatically. For considered purchases, they will act like consultants. Even big life events such as weddings, buying cars or homes, could involve an agent serving as coach and guide.
- Impulse, routine, considered, and life purchases: Different categories will see different patterns. Agents may nudge attention for impulse buys, quietly manage routine essentials, provide context-rich recommendations for lifestyle goods, and act as advisors for once-in-a-lifetime decisions. We’ve seen the early user-behaviour on the routine purchases with Alexa Routines.
- Voice commerce becoming viable: With agents handling context, permissions, and execution, voice commands can move beyond simple actions to nuanced requests. “Order the same sunscreen we used last summer, two bottles this time, from whichever store delivers fastest” will no longer be a stretch. Imagine talking to ChatGPT to explain what you want, chat with an AI agent shopping assistant that clarifies your preferences and checks out with the product at the end of the process.
- New competitive dynamics: General-purpose LLMs like ChatGPT or Claude, verticalised agents in niches like fashion, and new agent-native platforms could become dominant gateways. Platforms with structured data at scale — Amazon, Walmart, Shopify — are well positioned. Shopify in particular is embedding itself directly into the agentic ecosystem by powering OpenAI’s checkout. Independent brands will face a choice: integrate with these giants or form coalitions with unified APIs to stay discoverable.
From AI product discovery to checkout directly via an LLM Chatbot
To unlock this future, several layers need to evolve to make sure the infrastructure is ready for Agentic commerce:
- Structured data: Reliable product information, high quality pictures and videos and verified reviews.
- Unified APIs: Seamless agent access across retail platforms for carts, checkouts, and order management.
- Identity and memory: Persistent profiles that capture changing preferences, thresholds, and past purchases.
- Embedded capture: Collecting intent data within the user journey to improve recommendations.
- Agent-native infrastructure: Payments, logistics, and metadata (delivery windows, return policies, sustainability impact) exposed in ways agents can interpret.
In essence, AI agents shift ecommerce from “frictionless checkout” to “frictionless demand.” Intent is captured and executed ambiently, platforms behave like APIs, brands are treated like metadata, and the interface fades into the background.
While some incumbents like Shopify have been leaning hard on AI, most of the ecommerce industry is only now starting to experiment with these changes. The companies that build for this agent-mediated future, focusing on speed, data quality, and deep integration with user needs, will be the ones to define the next era of commerce. Our investment in Aigency AI (still in stealth) is an example of this new breed of Agentic AI for commerce leaders.
I'm a former general partner at google ventures, head of Google for Entrepreneurs in Europe and founding head of Campus London, Google's first physical hub for startups.
I'm also the founder of Techbikers, a non-profit bringing together the startup ecosystem on cycling challenges in support of Room to Read. Since inception in 2012 we've built 11 schools and 50 libraries in the developing world.
- Remagine Ventures Ranked Among Israel’s Most Active VCs in 2025 - March 18, 2026
- The Global Race for AI Sovereignty: Where Does Israel Stand? - March 16, 2026
- The Web’s Next Customer Isn’t Human - March 13, 2026
