AI Agents in E-commerce: Google’s New Protocol and the Future of Digital Commerce

A structural shift in digital commerce

The recent announcement of a new open protocol designed to facilitate commerce through artificial intelligence agents represents a structural shift in the e-commerce ecosystem. This is not a simple performance improvement or an isolated feature, but a fundamental redefinition of how sales platforms, consumers and automated systems communicate with one another.

This protocol is designed to create a common language between AI agents and e-commerce platforms, enabling complete purchases to be carried out without the need for traditional web navigation. Users no longer interact directly with websites and instead delegate that interaction to intelligent assistants capable of acting autonomously. This concept, commonly referred to as agentic commerce, marks the beginning of a new chapter in digital commerce.

The goal is to eliminate technical friction and create an interoperable ecosystem in which agents, online stores and payment systems can cooperate in a standardised way under a new global network protocol dedicated to AI agents, without directly impacting what e-commerce platforms expose to the web via HTTP, namely the HTML experience designed for human users.

The challenges e-commerce platforms face under this new protocol

Despite its vast potential, this new paradigm presents significant challenges for existing e-commerce platforms. Most current systems were designed for direct human interaction, with visual interfaces, navigation flows, product pages and traditional conversion funnels. The new protocol requires a radically different approach.

Key challenges include:

• Exposing commercial data in a structured, standardised and unambiguous way
• Responding to automated transactional requests in real time
• Managing authentication and permissions for AI agents
• Ensuring security in transactions executed without direct human involvement
• Maintaining brand consistency outside the website’s visual context

Platforms that fail to adapt their technical architecture risk becoming invisible to AI agents. In a scenario where assistants autonomously decide where to buy, being technically inaccessible is equivalent to not existing. In this context, investing in systems and infrastructure is no longer an operational concern but a strategic decision.

How the new protocol improves access to information on sales platforms

One of the most significant advances of this protocol lies in how AI agents access commercial information. Today, many agents rely on HTML scraping, simulating human browsing behaviour. This approach is inefficient, error-prone and often restricted for security reasons.

With the new protocol, agents no longer interpret visual pages and instead consume structured information via dedicated interfaces. This enables:

• Direct access to up-to-date product catalogues
• Real-time queries of pricing, stock and variants
• Clear interpretation of commercial terms and return policies
• Objective comparison across different platforms

This shift removes ambiguity, reduces interpretation errors and significantly improves the quality of decisions made by agents. E-commerce evolves into a set of transactional services consumed programmatically, aligning with modern, highly scalable architectures.

Transactions without navigation: cart, checkout and wishlists managed by AI

Another transformative aspect of this protocol is the ability to complete full transactions without any web navigation. AI agents can add products to carts, manage wishlists, apply promotions and execute checkout processes entirely autonomously.

For users, this delivers a seamless and continuous experience integrated into their virtual assistant. For platforms, it requires a complete rethink of transactional flows, which shift from page-based sequences to logical operations executed by external systems.

This model enables scenarios such as:

• Automated recurring purchases
• Cross-platform shopping list management
• Invisible and instant checkouts
• Integration between personal and business purchases

The challenge is no longer converting human users, but ensuring that agents correctly understand the platform’s value proposition.

The positive impact on bot proliferation and usability metrics

One of the less obvious yet highly significant benefits of this new protocol is its impact on the current issue of excessive AI agent bots crawling the web. Many e-commerce platforms today face:

• Artificially inflated traffic
• Distorted usability and conversion metrics
• Increased infrastructure costs
• Difficulty analysing genuine user behaviour

By introducing an official, authenticated and structured channel for AI agents, there is no longer a need for them to mimic human browsing. This dramatically reduces data noise and restores more reliable insights into real consumer behaviour.

It also enables new approaches to data analysis and analytics, focused specifically on agent–platform interactions.

A new era of AI-assisted purchasing

This protocol also marks the beginning of a new era of purchasing mediated by AI virtual assistants. Users stop manually searching for products and instead express intentions, preferences and constraints, delegating execution to their assistant.

This model enables:

• Context-aware and history-based purchases
• Decisions aligned with budget and personal values
• Reduced cognitive load during decision-making
• Greater efficiency in frequent or complex purchases

For brands, this represents a profound shift in how value is communicated. Visual persuasion gives way to data clarity, algorithmic trust and informational consistency. Digital transformation strategy must now account for both human users and artificial decision-makers.

Blockchain as a trust layer for AI agent transactions

Although still speculative, blockchain may emerge as a natural trust layer within this new automated commerce ecosystem. In a scenario where AI agents execute transactions autonomously, transparency, auditability and trust become critical.

Blockchain could enable:

• Immutable records of agent-executed transactions
• Automated and verifiable purchase auditing
• Smart contract execution for returns and warranties
• Decentralised identity for agents and platforms

This approach may be particularly relevant in B2B environments, global marketplaces or regulated industries, where proof of execution and accountability are essential.

Preparing today for tomorrow’s e-commerce

The new protocol for AI agent commerce is not merely a technological innovation. It is a clear signal that e-commerce is evolving towards a more automated, more efficient and less navigation-dependent model.

Companies that invest early in technical architecture, interoperability, structured data and strategic vision will be best positioned to lead this next phase of digital commerce. Agentic commerce is not a distant concept. It is starting now.

Preparing for this future is a strategic decision that will separate relevant platforms from those left behind.