A Tale of Two Commerce Protocols

In previous posts I discussed the advent of Agentic Commerce and how that is primed to become the new way to shop for products online.

In order to enable the AI platforms to be aware of your brand presence and product information there are a number of strategies and techniques, specifically GEO (Generic Engine Optimization) and AEO (Answer Engine Optimization), that can attract the AI bots to prefer your brand and recommend your products within the many conversations that customers are now having with AI applications.

GEO is a broad strategy that involves a number of techniques that involve changes in how you write product content and optimize your websites so that AI will pick you first as the authoritative source for the answers within the Agentic Commerce experience.

Very recently a couple of new developments have emerged that both sound like it’s attempting to answer a similar question. Namely OpenAI’s Agentic Commerce Protocol (or ACP) and Google’s Universal Commerce Protocol (or UCP).

OpenAI’s ACP is an open, cross-platform protocol designed to enable shopping and payments directly within AI assistants, independent of any single platform or user interface. It allows AI agents to discover products via merchant-provided feeds, surface accurate pricing and availability, and autonomously initiate checkouts on the user's behalf without redirecting them to an external website.

The checkout process uses secure, delegated payment tokens (which are single-use, time-bound, and amount-restricted), while ensuring that the merchant retains full control over settlement, refunds, chargebacks, and compliance. The first implementation of this protocol is the Instant Checkout experience within ChatGPT.

Google’s UCP is a new open standard designed to establish a common language that allows AI agents, businesses, and payment providers to work together across the entire shopping journey from product discovery to post-purchase support. They also have massive Industry Endorsement collaborating with the likes of Etsy, Shopify, Best Buy and Walmart (US) who are either implementing, or have gone live with AI Agents.

While it is designed to be compatible with other agentic protocols, UCP is initially rolling out exclusively on Google-owned surfaces, such as Search AI Mode, Google Shopping, and the Gemini App. It enables shoppers to buy from eligible retailers directly during product discovery without leaving Google, utilizing Google Pay for seamless transactions while the retailer remains the seller of record.

Why ACP/UCP are More Helpful Than AIO/GEO

While Artificial Intelligence Optimization (AIO) and Generative Engine Optimization (GEO) are critical strategies, they are fundamentally focused on top-of-funnel visibility. AIO and GEO ensure that an AI model correctly parses, embeds, and cites your brand as the "source material" when answering a user's question. However, simply getting found is only the first step of the commerce journey.

ACP and UCP are arguably more helpful because they bridge the gap between discovery and execution, transforming the entire commercial funnel:

• Moving from Recommendation to Action: AIO/GEO might prompt an AI to recommend your product, but the user still has to navigate to your site, browse, add to cart, and manually checkout. ACP and UCP grant the AI "agency" to act on the user's intent and execute the purchase directly within the conversational interface.

• Frictionless Shopping: Traditional e-commerce is linear and rigid (search → browse → filter → product page → cart → checkout). ACP and UCP collapse these steps into a natural dialogue, drastically reducing friction and lowering cart abandonment.

• Capturing Immediate Revenue: By allowing shoppers to move from intent to purchase without breaking context or leaving the app, these protocols turn high-intent discovery moments directly into revenue.

In short, AIO and GEO help AI talk about your product, but ACP and UCP allow AI to buy your product on the customer's behalf.

Which one to choose?

Both OpenAI's Agentic Commerce Protocol (ACP) and Google's Universal Commerce Protocol (UCP) share the same overarching goal: to reduce friction in the shopping journey by allowing AI agents to handle product discovery and checkout seamlessly, without redirecting the user to an external website.

However, they differ significantly in their execution environments, how they handle payments, and their initial scope.

OpenAI’s Agentic Commerce Protocol (ACP)

• Design & Environment: ACP is an open, cross-platform protocol built to enable shopping and payments directly within AI assistants. It is designed to be independent of any single platform, user interface, or distribution surface. Currently, its primary implementation is the "Instant Checkout" experience inside ChatGPT.

• Payment Mechanism: ACP initiates checkout on the user's behalf using delegated payment tokens. These tokens are highly secure because they are single-use, time-bound, and amount-restricted.

• Merchant Role: In this model, merchants maintain complete control over the transactional backend, retaining responsibility for settlement, refunds, chargebacks, and compliance.

Google’s Universal Commerce Protocol (UCP)

• Design & Environment: UCP is pitched as a new open standard designed to support the entire shopping lifecycle, from product discovery and buying to post-purchase support. However, unlike ACP's cross-platform focus, UCP is initially being rolled out exclusively across Google-owned surfaces, including Search AI Mode, Google Shopping, and the Gemini App.

• Payment Mechanism: Instead of delegated tokens, UCP leverages Google Pay to complete transactions natively during product discovery, with PayPal support planned for the future.

• Additional Features: Alongside UCP, Google launched a feature called "Business Agent," which allows retailers to engage shoppers conversationally and enable direct purchases right within Google Search.

The Core Differences

• Where the Shopping Happens: ACP enables agent-led commerce primarily across the OpenAI ecosystem as a standalone destination, while UCP currently focuses on reducing checkout friction specifically within Google's massive search and discovery surfaces.

• Coexistence Over Competition: Google designed UCP to be compatible with other agent-to-agent standards and protocols. This means the two protocols are not necessarily meant to replace one another, but rather to coexist. UCP helps convert high-intent shoppers who are actively searching on Google, while ACP opens the door to new demand where AI chat assistants act as the shopping destination.

So it’s not like the old VHS/Beta video wars of the 80s. The question isn't which protocol "wins" it's whether your product data (and infrastructure) is ready to feed both. The reality is that you may need to support a multi-protocol ecosystem, just like supporting Apple Pay, Google Pay, and PayPal today. We are entering a multi-agent, multi-protocol world where structured product data is the "source code" of commerce.

Getting to grips with Generative Engine Optimization (GEO)

In the next few posts I am going to break down several new (new to me) concepts in the world of AI, that are especially important to anyone attempting to sell products online. In my last post I set out the main features of the Agentic Commerce landscape, however in order to break into that world you need to understand how that can actually be done. And the first concept is a content strategy called Generative Engine Optimization (GEO).

In the traditional world of promoting your brand or products you used to fight tooth and nail to be the top blue link on Google by employing SEO (Search Engine Optimization) tactics to your content. Well, the game has changed. If SEO is asking to be put on stage, GEO is asking to be the script the speaker reads from. So, in this post I’ll try and lay out the basics on what GEO is and how you can actually win at it without needing a PhD in computer science.

What on Earth is GEO?

Generative Engine Optimization is the art of convincing AI engines, like ChatGPT, Perplexity, and Google’s AI Overviews, that you are the most trustworthy, relevant source to mention when they build an answer for a user.

In the good old days (which is now reaching back as far as 2023), you wanted a click. Now, you want "influence." You want the AI to read your content, understand it, and synthesize it into its answer, hopefully citing you as a source. The goal isn't just ranking; it's getting recommended by a machine that acts like a smart friend giving advice.

Here is the cheat sheet on how GEO shakes up the old SEO playbook.

1. The Goal: Clicks vs. Influence

• SEO (Old School): The goal is to rank your website at the top of a list of "blue links" so a human clicks on it and visits your site.

• GEO (New School): The goal is to get your content read, understood, and synthesized by the AI into the answer itself. You aren't just trying to get a visit; you are trying to be the source the AI quotes to sound smart.

2. The Tactics: Keywords vs. Credibility

• SEO: You focus on keywords, backlinks, and meta tags to tell the search engine what your page is about.

• GEO: You focus on "citation-worthiness." This means using hard statistics, direct quotes, and authoritative facts so the AI trusts you enough to build its script around your content.

3. The Result: A List vs. A Conversation

• SEO: Produces a directory of options for the user to sift through.

• GEO: Produces a single, conversational answer that solves the user's problem immediately—often without them ever needing to leave the search interface (a "zero-click" interaction).

The Playbook: How to Win at GEO

You can’t just stuff keywords into a page and pray anymore. You have to create content that AI loves to read and summarize. Here is how you accomplish that.

1. Be the "Comparison" Expert

AI engines love synthesis. If a user asks, "Milwaukee Power Drill vs. Makita Power Drill" the AI looks for content that weighs the pros and cons so it can build an answer.

• The Move: Don’t just talk about yourself. Create honest comparison pages, "best of" lists, and detailed reviews.

• Example: Instead of a page that just says "We have a wide range of power drills" build a page comparing your drill selection against pricing, performance, specifications such as actual battery life and torque. If you do the heavy lifting of comparing options, the AI is more likely to use your analysis.

2. Feed the AI Facts and Stats (It Loves Data)

Generative engines are hungry for authority. They trust numbers more than vague marketing fluff.

• The Move: Add citations, quotations, and specific statistics to your content.

• Example: Don't write, "The power drill packs a lot of torque" That’s fluff.

• Better: Write, “Engineered with a POWERSTATE™ Brushless Motor that hammers out a massive 1,400 in-lbs of torque….”.

• The Result: Research shows that just adding citations can significantly boost your visibility in AI answers. If you can find an independent article like a scientific journal explaining why torque is important for drilling, it shows that you have done research and that your claim is grounded in reality.

3. Structure Your Content Like a 5-Year-Old Needs to Read It

We will look deeper into optimizing your pages to better suit AI scanning in future posts, and also the use of Schema Markup to help that layout. If your website is a wall of text, the AI might skip it. It needs structure to parse information quickly 8.

• The Move: Use clear headings, bullet points, and tables. Break it down.

• Example: If you are selling a product, use a table to list benefits and features rather than burying them in a paragraph. This makes it incredibly easy for the AI to scan your page and say, "Aha! Here is the answer," and pull that data into its response.

3. The "Inverted Pyramid" (Answer First)

AI is impatient. When scanning for an answer to cite, it prioritizes information found at the top of the section. If you bury the lead, you lose the citation. This is sometimes called "Position-Adjusted Visibility".

• The Goal: Answer the user’s question immediately, then explain the details.

• Example:

• Scenario: You are an HVAC company writing about "Why is my AC blowing warm air?"

• The Move: Do not start with a story about summer heat. Start with a 40–60 word paragraph that lists the top three reasons (e.g., dirty filter, low refrigerant, frozen coils).

• Why it works: LLMs (Large Language Models) grab that clean, concise paragraph to form their summary. If you make it easy for them to "steal" your summary, they will cite you as the source.

4. Get the "Crowd" to Back You Up (Citations Everywhere)

This is the part most businesses miss. AI doesn't just trust you; it trusts what everyone else says about you. This is the "Corroboration & Authority" principle. If your website says you are the best, that's marketing. If Reddit, Yelp, and a "Top 10" listicle say you are the best, that's a fact.

• The Move: You need to be mentioned in places other than your own site. This includes Reddit threads, "Top 10" listicles in your niche, and review sites like G2 or Yelp.

• Example: If someone asks an AI for the "which is the best percussion drill for industrial use?" the AI checks listicles and directories. If you aren't on those third-party lists, you likely won't get recommended, even if your website is beautiful.

• Why it works: When an AI constructs an answer, it looks for consensus. If you appear across multiple authoritative sources, the AI feels "safe" recommending you,.

5. Leverage "Agentic" Content

This ties into a topic for my next post. We are heading toward "Agentic Commerce," where AI agents do the shopping for people. To get picked, you need to answer specific questions which follow a subtly different strategy of Answer Engine Optimization (AEO).

• The Move: Create "Question/Answer" pairs and detailed use cases.

• Example: Instead of a generic product title like "Cordless Drill," broaden your description to answer the inevitable questions: "This Cordless Percussion Drill can drill into Wood, Concrete, and Steel with a battery life of 8 hours". You are proactively answering "What can it drill?" and "How long does it last?" before the user even asks.

The Bottom Line

GEO is about making it easy for the machine to trust you, your brand and your products. If you are transparent, fact-based, widely cited across the web, and structured clearly, the AI will reward you by making you part of the conversation. Don't just be a link - be the answer.

What is Agentic Commerce?

What Agentic Commerce Means for the Regular Consumer

My day job involves working for a software company that develops a PIM (Product Information Management) solution. Maybe I will reveal what company that is in future posts, but I'd like to keep this fairly generic and sales free, and that way I may learn a thing or two about any other competitive systems too. The point is that this blog will possibly start off with a bit of a skew toward product content, selling products online and how AI is helping in that process. But I’m just going to meander and see what I uncover and hopefully it might be interesting to someone…

The first big topic that everyone is talking about (in my world) is Agentic Commerce and for the average shopper this represents a shift from "Do It Yourself" to "Do It For Me."

In the current world of e-commerce, the burden of work falls on you. You have to think of keywords, type them into a search engine, open ten different tabs to compare prices, read through reviews to check for quality, manually filter for your size, and finally click through a checkout process.

Agentic Commerce changes this dynamic by introducing AI Agents. Unlike standard chatbots that just answer questions, these Agents have "agency"—the ability to reason, plan, and execute tasks to achieve a goal you set for them.

Here is what this looks like in practice for a consumer:

1. The "Personal Shopper" Experience Imagine you need a gift. Instead of browsing endless categories, you simply tell your AI Agent: "Find a birthday gift for my brother, under $50, something tech-related".

• Reasoning: The Agent understands the constraints (budget, interest, occasion). You may already have told it your shoe size and waist measurement to streamline the next step.
• Planning: It scans catalogs, checks reviews for reliability, and compares prices across different sellers instantly.
• Action: It presents you with the best options—or, if you trust it enough, simply buys the best one for you.

2. Complex Tasks Handled Instantly Agents shine when the request is complicated. If you say, "Book me a flight to Barcelona for Friday morning," the Agent doesn't just show you a list of links. It autonomously compares airlines, finds the best deal that includes baggage, books the ticket, and emails you the confirmation. It acts like a knowledgeable consultant rather than a search bar. It may even handle the rest of the experience by organizing your boarding pass, alerting you to any delays and later managing any insurance claims should they unfortunately arise.

3. Running Errands on Autopilot For repeat purchases, Agentic Commerce offers "set it and forget it" convenience. An Agent can notice when you usually buy cleaning supplies or cosmetics and automatically reorder them so you never run out, handling the transaction in the background.

4. The End of "Tab Fatigue" The most immediate benefit for a consumer is the removal of friction. Because the Agent can read product specifications, check inventory, and understand shipping policies instantly, you no longer need to hunt for this information. The Agent digests it all and provides a direct solution, making the actual act of shopping nearly invisible but highly effective.

It's happening now!

Well, kind of. While we are not yet seeing this agentic concept fully realised, I can assure you it is starting to emerge, at least in the US, where the waves are lapping at the shore.

As far as infrastructure is concerned, OpenAI and Google have produced specifications to enabled sellers to upload their product information. Stripe, Klarna and Mastercard are at the forefront of supporting the transactional side of things to allow machine-to-machine payments, effectively making the shopping cart part of the agentic experience.

Several retail giants (customer-facing agents) have integrated capabilities directly into their apps to move beyond basic search:

• Amazon (Rufus & "Buy For Me") : As of 2026, Rufus has evolved to include a "Buy For Me" feature. It can monitor prices and autonomously execute a purchase when a specific price threshold is met. It can also digitize a handwritten grocery list and populate a cart instantly.
• Walmart : Using the OpenAI protocol, Walmart allows users to build complex carts (multi-item transactions) through conversational agents that understand "standing intents" which are pre-defined, persistent user goals or purposes that an AI system is trained to recognize and respond to, for example "Always keep my pantry stocked with organic milk")
• Shopify (Sidekick & Agentic Storefronts) : In early 2026, Shopify launched "The Renaissance" update. Sidekick is now a proactive agent that manages merchant operations (inventory and promo generation) and distributes product data to third-party agents like ChatGPT or Perplexity for "headless" agentic checkout.

Specialized AI platforms or "agent-first" companies are where many consumers are now starting their shopping journeys:

• Perplexity (Buy with Pro) : Launched in late 2024 and expanded in 2025/2026, this tool allows Pro users to research a product and click a single button to have Perplexity's agent handle the checkout process on a merchant's site.
• Microsoft (Copilot Checkout) : Integrated into the Microsoft 365 and Edge ecosystem, this allows agents to "negotiate" with brand agents and secure delivery slots that fit the user's Outlook calendar as well as complete the shopping cart within the customer conversation.

And finally, many software vendors are implementing solutions to help B2B/B2C enterprises manage their own agentic selling experiences and deliver product information to it:

• Salesforce (Agentforce) : Actively used by B2B and B2C enterprises to automate order management, returns and exchanges. Salesforce agents can now autonomously process refunds and update ERP systems based on a customer's conversation.
• UiPath ("Autopilot" and "Maestro") : These platforms allow enterprises to build custom agents that handle procurement, invoice matching and automated restocking without human intervention.
• Agility PIM : Using its built-in AI Content Generation to upscale creation of answer-optimized product information that is now required to tackle new content strategies such as GEO and AEO.

What does this mean for online sellers?

So this leads to a final point about what it means for you to sell products online, you can no longer just rely on attracting customers to your websites, or even other affiliated or distributor sites. Traditionally this process was done by optimising your content through SEO (Search Engine Optimisation) that sought to rank your product pages as high up the Google search results as possible. But in an Agentic Commerce world your customers may never even set eyes on your product’s pages. So you must seek to be part of the AI conversation rather than a link to your product.

So my journey into this world will cover a bunch of topics that will try to unravel big topics that sellers will need to be aware of as they make this shift (I nearly said "paradigm shift" but I really need to avoid such cliches, they are not funny).

• Generative Engine Optimization (GEO & AEO) : Sellers can no longer just optimize for Google keywords (SEO). They must learn AEO (Answer Engine Optimization) and GEO, strategies focused on becoming the "source material" or "cited answer" for AI models like ChatGPT and Perplexity,. This involves structuring content so an AI can easily read, summarize, and recommend it.
• The New "Product Feed" Standard: To sell to an AI Agent, your data must be flawless. OpenAI and Google have released their own specific product feed specifications that require high-frequency updates and new data fields like "Mood" or specific use cases, ensuring the AI understands context, not just specs.
• Agentic AI PIM (Product Information Management): Preparing data manually for AI Agents is impossible at scale. Sellers should explore Agentic AI PIM, where AI tools autonomously clean, enrich, and structure product data to ensure it is trusted by shopping bots.
• Trust and Transparency Signals: In an era where a machine buys on behalf of a human, trust is currency. Sellers will need to explore how to present "Seller Policies" and transparency data (like return rates and privacy policies) directly to AI Agents to ensure they are selected as a "safe" purchase option.