Pattern Pi turns AI shopping rankings into an ecommerce control loop

Pattern's system has already recovered millions of featured offers, adjusted product pricing, and fixed content listings across the brands it manages — without human review. The company calls this autonomous execution. What it describes in its own words is a control loop: a system that watches how a product ranks inside Alexa for Shopping or ChatGPT Shopping, then automatically changes the price or content to move that ranking, around the clock.

"The loop closes," said one person familiar with Pi's design, who asked not to be named because the system is not yet publicly documented in full. "It's not A/B testing. It's not marketplace bidding. It's a control system."

The claim is that Pattern has built the first commercial system treating AI shopping agents as ranked distribution endpoints to be actively optimized — not just monitored. The Q1 numbers are the closest thing to proof: Pattern reported $774 million in revenue in the first quarter, up 43 percent from a year earlier, and $54 million in adjusted EBITDA, up 59 percent. Whether those numbers validate the control loop or simply reflect a broader ecommerce tail wind is the question this story sets out to answer.

The scale of what Pattern is claiming is not small. Non-Amazon revenue grew 119 percent year over year; international revenue grew 101 percent. The company says Pi has already taken millions of automated actions — recovering featured offers, adjusting prices, fixing product content — since its deployment across Pattern's brand portfolio, per the press release.

What Pi tracks is notable: the system monitors how products rank with four major AI shopping agents — Amazon's Alexa for Shopping, Walmart's Sparky, ChatGPT Shopping, and Google AI Mode, according to the press release. Those four agents represent a meaningful slice of how consumers will discover products as AI reshapes search. Morgan Stanley projects nearly half of online shoppers will use AI shopping agents by 2030, accounting for roughly 25 percent of spending. If that projection holds, the ability to control how products rank inside those agents becomes as structural a cost as Amazon advertising is today.

The differentiation Pattern is claiming is the closed loop: most brands can see how they rank in AI shopping results. Pi is designed to both see and act — automatically. The company cites 77 trillion data points accumulated over 13 years of operating on behalf of global brands, growing by more than 800 billion new points weekly, per the press release. It holds 41 patents related to its technology portfolio.

"AI agents aren't a future interface — they're a new operating layer for commerce," said Ryan Byrd, Pattern's chief technology officer, in research the company published in January. "Brands that treat agentic AI like just another marketing channel will fall behind."

Pattern's own numbers lend some weight to the urgency. In a survey of 1,000 senior business leaders across the US, UK, Germany, and the UAE that Pattern conducted, 76 percent of ecommerce organizations reported reduced customer acquisition costs as consumers began relying on AI-driven tools for product discovery, per the RTIH report. One in three already deployed AI shopping agents. Eighty-seven percent expected AI-powered search to drive direct sales growth over the following 12 months.

The counterargument is straightforward: Pattern has not published peer-reviewed benchmarks for Pi's performance, nor has it disclosed the specific mechanism by which the control loop closes. The 41 patents and 77 trillion data points are verifiable facts; the claim that they constitute a defensible moat is an assertion. No brand customer has been quoted publicly on Pi's specific results. Pattern declined to make a brand client available for this article.

What the company has said publicly is consistent: Pi is an autonomous execution engine built on operational history, designed to act on signals from AI agents the same way a trader acts on price signals — automatically, continuously, and at scale.

The question is whether that control loop is real or rhetorical. If it is real, Pattern has built something structurally new: a toll collector positioned between brands and AI-driven commerce, collecting for the privilege of being optimized. If it is marketing language over incremental automation, the story is a product announcement dressed in AI vocabulary.

Pattern's Q1 financials, filed with the SEC on May 6, will be the first test. Revenue growth of 43 percent and EBITDA growth of 59 percent are the kind of numbers that either confirm the model is working or set up a harder fall if the conversion metrics plateau. The 17-to-19 percent conversion rate improvement Pattern cited — from its AI investments broadly, per PYMNTS — is directional but not proof of a closed loop. That distinction matters.

What is clear is that the race to optimize for AI shopping agents has moved from theory to commercial operation. Whether Pattern's loop actually closes, or whether it is the first to claim it does, will become apparent in the next earnings cycle. Brands watching this space should watch what Pattern's clients say after Q2 — not what Pattern says in its press releases.

Pattern, for its part, appears confident enough to go first.