The $8 Trillion Shift to AI Agentic Commerce

From Search To AI Agents

The way we interact with the Internet has changed over time, from the early websites to smartphone apps and social media.

But fundamentally, the premise of e-commerce and the online economy has been relatively consistent: ads, search, or social media bring traffic to a website/app, and the customers browse until they find something of value to them, to buy/subscribe/follow/etc.

This might change soon with the rise of AI. Instead of just relying on an automated system for discovery, people might soon rely on AI agents to perform tasks for them, including online purchases, completely changing our approach to digital interactions.

Source: ARK Invest

This is going to be especially revolutionary due to the extremely quick speed of adoption of AI tools, much quicker than the speed at which PC users adopted the Internet.

Source: ARK Invest

Behind these potential changes is the quick improvement of AI performance, declining costs for tokens, and exploding investment in new data centers to provide the infrastructure to the AI industry.

The End of Search-Driven E-Commerce

So far, online shopping has relied on the “Search & Click” approach, with most of the changes in the past 30 years being how the search task is performed, from early Internet indexes to search engine like Google (GOOGL -1.16%), to e-commerce platform like Amazon (AMZN -1.79%) and then to social media platform like Facebook and Instagram (META -2.65%).

Source: ARK Invest

Through the evolution of e-commerce, the process has become increasingly seamless, with less and less time spent before a purchase. Now, the mere minutes taken for a mobile purchase are getting even shorter, with AI agent-driven purchases taking as little as 90 seconds.

Source: ARK Invest

The idea is that instead of browsing a website or reacting to ads, users can send a customized AI agent, finely tuned to their preferences, to look for a specific product for them.

They let the AI perform the entirety of the discovery and engagement process, and all but the last step of decision making leading to the purchase, with the final human user only involved in validating the AI suggestion at the very end.

Source: ARK Invest

The investment firm ARK Invest estimates that AI agent-driven purchases could grow from 2% of online spend in 2025 to ~25% by 2030. This could represent up to $8T of online consumption, so even a fraction of this value captured by the AI agents would be in the tens or hundreds of billions of dollars.

Source: ARK Invest

The Agentic Commerce Protocol Stack

This evolution toward AI agents will require a key evolution of the underlying infrastructure of e-commerce websites, so that agents can efficiently understand them.

Source: ARK Invest

MCP

The first part is the Model Context Protocol (MCP), an open-source standard introduced by Anthropic in late 2024.

MCP can let the agent perform real-time, multi-step actions without requiring custom-built integrations. So it can let an e-commerce website have agents that manage inventory, update product listings, and process orders in real time.

ACP

The Agentic Commerce Protocol (ACP) is an AI open standard for communication and collaboration between different AI agents.

The idea is to facilitate the creation of hierarchical, multi-agent workflows, with humans in the decision loop at key points of the whole process, but with AI agents managing the rest either independently or collaboratively.

As such, ACP will be important for securing autonomous payments between the agents used by both the buyers and sellers.

From SEO to AEO (Agent Engine Optimization)

Swipe to scroll →

To some extent, brand loyalty and other marketing tools centered on a company have already been replaced by algorithm-driven processes like ranking in search, ad-driven traffic, e-commerce platforms’ internal ranking & keywords, etc.

This will keep evolving with the arrival of agents, as SEO (Search Engine Optimization) is partially giving way for AEO (Agent Engine Optimization). While they both will have some similarity, like the importance of machine readability for the searching “bot”, be it from a search engine or an agent, there are also some key differences:

• AEO must enable agents to complete tasks, such as booking services or fetching prices via API, through clear, direct data.
• Content must move from keyword-heavy, conversational text to direct, factual answers.
• AIs are more reliant and trusting of verified sources, citations, and high-quality, trustworthy content to recommend brands to their users.
• At least as long as AI is booming and changing quickly, AEO requires optimization across various AI platforms and browsers like ChatGPT, Claude, Gemini, and Perplexity, while in most countries, SEO usually really means “optimized for Google”.
• Even more than with SEO, data transparency, as well as API reliability, will be rewarded by AI agents.

Impacts On E-Commerce Ecosystem

The shift to AI agents is important, as AI search is already becoming a significant segment of the search market, with it making more than half of search traffic expected before 2030.

AI Search ads are expected to grow too, although at a slower pace, reaching “only” $200B spent in 2030.

Source: ARK Invest

Overall, it might favor, even further than today, websites built with advanced tools following the right agent-focused protocols and standards.

However, as agents will care more about the product than the graphic user interface or ease and comfort to use the same website, or human habits, this could genuinely hurt platforms that do not make enough effort to cater to agents, and instead help independent, stand-alone websites.

Maybe more unexpectedly, agents might bring back a focus on the real quality of products, as it will be very hard to create genuinely looking fake reviews and the agents will likely be more able to identify if their users are a real person or not, and ignore such reviews in their evaluation.

Economic Impact: The AI Productivity Surplus

AI agents for online purchase are just one segment of the impact AI agents will have. AI is expected to turbocharge the productivity of workers in fields like translation, law, scientific research, art, healthcare, security, logistics, transportation, finance, etc.

The result should be a rising productivity, depending on the speed of adoption and the scale of investments made in AI technology, with software expected to go 2x-10x by 2030, generating anywhere between $3.4T to $13T in surplus productivity.

Source: ARK Invest

This will likely cap knowledge worker employment growth and hours worked, up to 81% of tasks.

In turn, it will also free up time for high-level agent management, be it in a professional capacity or as a consumer.

Public Companies Positioned for Agentic Commerce

Shopify (SHOP)

Shopify was founded in 2004 as a tool to build e-commerce websites. To fuel growth, Shopify launched its application programming interface (API) platform in 2009, allowing for the creation of applications.

These applications added vital features such as website themes and streamlined refund processes, making Shopify one of the most customizable e-commerce platforms in the world. The company grew rapidly over the next several years before eventually going public in 2016.

The company is continuing its tendency to move ahead of e-commerce trends thanks to its Universal Commerce Protocol, launched together with Google. With it, Shopify aims to become an infrastructure layer for commerce using AI agents.

UCP was co-developed by Google and Shopify with support from major commerce and marketplace players, including Etsy, Wayfair, Target, Walmart, and others.

Source: GeekSeller

It is designed to make it fast for agents built within Google AI Mode, Gemini, Microsoft Copilot, ChatGPT, and others to handle the messy reality of commerce: discounts, loyalty programs, subscriptions, pre-orders, and edge cases in checkout.

The company is also moving fast to expand into this new field, notably by opening its Shopify Catalog, a normalized, enriched catalog of billions of products to brands that do not run their online store on Shopify.

It also launched a new Agentic plan for these brands to sell through AI channels, the Shop app, and future partners without migrating their storefront.

A key idea behind UCP is neutrality. The protocol is vendor-agnostic and not limited to Google or Shopify products, making it possible to support agentic commerce on virtually any platform.

UCP will directly compete with OpenAI’s Agentic Commerce Protocol, developed with Stripe and also released as open source. UCP aims to focus mostly on a standardized, end-to-end commerce layer that can operate at marketplace scale instead of agent behavior.

Besides AI efforts to move ahead of the AI agent revolutionizing e-commerce, Shopify is still quickly growing, with, for example, for Q3 2025, “GMV up 32%, revenue up 32%, free cash flow margin at 18%”.

As such, no matter the speed of adoption of AI agents, it is likely that Shopify will stay at the forefront of e-commerce, as both the main alternative to listing products on e-commerce platforms and the key partner of these platforms to adapt to agentic AI.

Latest Shopify (SHOP) Stock News and Developments