{ “@context”: “https://schema.org”, “@type”: “AnalysisNewsArticle”, “mainEntityOfPage”: { “@type”: “WebPage”, “@id”: “https://martech.org/the-real-risk-in-agentic-commerce/” }, “headline”: “The real risk in agent commerce”, “description”: “Agent commerce shifts purchasing decisions from loyalty-driven to brand human to algorithmic logic, creating massive exposure for retailers Brands are unprepared for cross-platform shopping bots Brands face loss of direct customer relationships and erosion of margins as autonomous AI agents prioritize price and specification matching over brand affinity. “sameAs”: “https://www.linkedin.com/in/anamourao/” }, “publisher”: { “@type”: “Organization”, “name”: “MarTech.org”, “url”: “https://martech.org”, “logo”: { “@type”: “ImageObject”, “url”: “https://martech.org/wp-content/themes/tdm-editorial/img/logo/logo.svg” } }, “abstract”: “The article assesses the operational and strategic vulnerabilities brands face in the transition to agent commerce, detailing how autonomous AI agents are disrupting traditional funnel dynamics, customer lifetime value, and direct audience relationships.”, “speakable”: { “@type”: “SpeakableSpecification”, “cssSelector”: ( “h1”, “.article-content p:first-of-type” ) } }
The conversation about agent commerce among marketing teams has narrowed down to a single question: how do we get AI to recommend us? It’s a reasonable question, but it’s not the right question to ask.
The brand an agent buys is one of the brands selected by the algorithm. A brand that a customer asks for by name is something a brand has to earn. As strategist Jess Graham According to him, the first case is purchasing, and no one has ever fallen in love with sourcing.
This distinction is important as agents play a larger role in product discovery. A mark can be perfectly readable by the machine and completely absent for the person. The agent evaluates, the customer receives a product and the act of choosing quietly disappears from the process.
For marketers who have martech and customer data, this is where the strategy can be tested. Algorithmic readability is important, as is remaining present to the human being behind the agent.
10X your SEO with Semrush for business.
The world’s most powerful SEO platform, built specifically for businesses.
@media (maximum width: 768 px) { .headline-responsive { font-size: 30px !important; line-height: 1.3 !important; } }
AI agents are already reshaping product discovery
Agent trading is already present on the market, but not in the form promised by the first titles. OpenAI Instant Payment went live in September 2025 and was withdrawn in March 2026five months later.
Only a dozen traders have integrated it; usage remained low and shoppers who searched in ChatGPT still preferred to complete the purchase on the retailer’s own site. OpenAI has not gone out of business. It has moved to a discovery-driven model that redirects shoppers to merchant apps and storefronts, where the retailer handles payment.
Google has taken the opposite path with its own protocol. THE Universal Trade Protocolannounced at NRF in January 2026 with Shopify, Etsy, Wayfair, Target, Walmart and Visa, is now live andcontinues to expand: Shopping cart support, catalog access and identity binding are already in use. Read the two stories together and one thing becomes clear. The payment layer is still taking shape, but agents are already in the discovery layer.
Consumer research shows that AI is already influencing product discovery. Salesforce found 39% of consumers, including 54% of Gen Z, have used generative AI to discover and review products. Adobe also has traffic tracked from generative AI tools to retail sites.
What research does not yet demonstrate on the same scale are agents who purchase on behalf of consumers. Salesforce found that 63% of Gen Z want AI agents to make purchases for them, but this reflects stated interest rather than measured behavior.
Consider the moment this could create. A customer asks an AI agent to purchase moisturizer. The agent weighs a thousand options based on price, reviews and speed of delivery, and buys one. The customer has never seen the packaging, never read the brand story, and never compared it to the product they may have already used and loved. The agent made his choice and the client received the result.
This scene comes from Graham, who mapped out what happens to brands when agents start making purchasing decisions for people.
Being readable for AI is not enough
Graham draws a clear distinction between algorithmic readability and brand preference.
Algorithmic readability involves structuring product data so that the agent can read it, classify it and place a mark in the consideration set or, as it is expressed, be at the head of the algorithm. This is where the budgets are going, and the first returns are real. Brands that appear in AI-generated responses gain more trafficand this traffic tends to spend more time there. The work is important and must continue.
The other challenge is to remain present to the human behind the agent. Graham’s name for the state of failure is agentic invisibility. The commercial implication follows directly from this. A brand purchased without having been deliberately chosen by a human may have given up its pricing power.
To an agent optimizing pricing, ratings, and delivery, an undifferentiated brand reads like a commodity, and commodities compete on price until someone loses. Graham gives it a name that belongs to a P&L: the discovery feethe compounding cost of giving in is how customers find a brand and then decide it’s worth choosing again.
The Cost of Giving Up Control of Discovery
Other industries have already paid the price for having relinquished control of the discovery.
- Hotels have outsourced discovery to online travel agencies and now pay between 15% and 30% per booking while having almost no customer data.
- Musicians have turned discovery over to playlist algorithms and now earn fractions of a cent per stream, with the platform deciding who gets featured.
- Third-party sellers created demand in a market, then watched the market study what was selling and launch its own competing products.
The sequence repeats in a loop. A new intermediary (in this case, AI agents) provides some convenience. Brands accept more unfavorable economic conditions to maintain their access. The middleman captures the customer relationship and data, and brand differentiation comes down to feature comparison.
Agent commerce is the most comprehensive version of this model so far because the customer is not even present when the evaluation takes place.
Brand preference requires data that you control
at Graham’s prescription for brands is about creating discovery experiences that agents can’t capture and giving people a reason to choose your brand from the start. The martech stack plays a vital role because to gain preference, it is not enough to make a brand readable to an agent.
Most brands striving for readability may provide agents with incorrect data. Static product feeds and last-touch attribution models describe what a customer did and remain silent on why they did so. This is enough to be classified. Being preferred requires more.
An agent optimizing on a few structured domains treats each well-structured competitor as interchangeable. This instability shows up in the AI results.
SparkToro found less than 1% chance that the same brand appears in two identical AI queries. A brand flickers, present in one response and gone in the next, with no explanation offered to the marketer or customer.
Escaping agentic invisibility then becomes a data decision before it becomes a campaign decision. This requires zero, first-party data that captures preference and relationship signals – those that explain why a customer chose a brand – kept somewhere in a brand’s own martech stack.
In practice, this means owning the moment of discovery rather than borrowing from it (via third-party data or walled gardens, for example). This means capturing signals from the community and direct conversations, treating delivery and unboxing as relational rather than logistical data, and building a consented identity that persists whether or not an agent is in the middle.
A brand that has the discovery, direct relationship, and underlying data becomes the brand an agent can be tasked to search by name. This is the sustainable version of the high-end algorithm, and it runs on data controlled by the brand.
What Marketers Should Do Now
On the data side, the hard truth is that it is an architecture problem before it is a strategy problem. A team that can’t connect the tools it already has won’t suddenly pick up on customer preferences and signals. The agentic moment does not create this gap, but exposes and amplifies it while increasing the cost of leaving it open.
Start with an audit rather than a tool. Map out what zero and first-party customer data is actually captured and be honest about it. A data set that records what was purchased and almost nothing about why it was chosen is built to be readable and stops there.
Then, deliberately decide where relational data is built in places that belong exclusively to your brand: owned discovery, community, Graham-designated delivery timing, and a direct channel that the platform can’t freeze. Consider algorithmic readability as the minimum due to the machine, then invest beyond that.
Give a real number to the discovery tax this quarter. A board that hears “we win the deal and lose the relationship” can fund the fix faster than another scorecard.
Place ownership of agent data where the customer relationship already resides, within marketing, rather than letting it drift into a purely technical decision made downstream.
The trade agent will assign the transaction to the most machine-readable person. Whether this will also pass on the relationship is a martech and data decision being made. Marketers need to lead this decision rather than leaving it as the default.
The position The real risk of agent trading appeared first on MarTech.
