Accelerate26 Session Recap: The Launch of Pattern Intelligence (Pi)

Ecommerce doesn't reward the brands that react the fastest. It rewards the brands that have stopped reacting altogether. That was the premise Pattern Co-Founder and CEO Dave Wright brought to the Lattice Stage at Accelerate 26, and the session that followed made a strong case for why that shift is happening right now.

Here's what went down.

The problem with ecommerce at scale

Running a global ecommerce operation means managing an impossible number of moving parts. Ad types, content, pricing, inventory, marketplace rules that change constantly across dozens of platforms. Pattern manages 150,000 SKUs across hundreds of countries and languages. At that scale, even the best teams can't catch everything fast enough.

The data layer makes it harder, not easier. Dave walked through a live example of asking a leading AI tool the same question twice in two separate sessions: is top-of-funnel advertising the best way to generate new sales? The answer changed between sessions. No memory. No consistency. No underlying data to anchor the response.

That's the problem Pi was built to solve.

What Pi actually does

Pi is Pattern's autonomous ecommerce execution engine, built on 13 years of operational experience and more than 77 trillion proprietary data points, growing by 800 billion new points every week.

The way Dave framed it: ecommerce runs on a formula. Revenue equals traffic times conversion times price times availability. Pi knows every lever in that formula. It runs active sensors across featured offers, advertising, content, pricing, and inventory. When a sensor fires, Pi acts. When a decision needs brand judgment, it surfaces a curated action item for approval instead of moving on its own.

Since rolling out across Pattern's brand portfolio, Pi has already taken millions of automated actions, running 24/7.

A live look inside Pi

Dave walked through a live demo using Leatherman as the example brand, and it was one of the more compelling product demos of the day.

The Pi home screen surfaced a weekly sales summary, flagged notable drivers behind a 2.7% week-over-week revenue climb, and pulled in relevant news affecting the brand in real time. No searching required. Just context, served up and ready.

From there, Dave showed how Pi maps the full ecommerce formula for a brand. Page views. Conversion. Buy box percentage. In-stock rates. Each metric clickable, each one connected to the underlying data and explainable through a live chat interface trained on Pattern's proprietary knowledge base and the brand's own guidelines.

One standout moment: Pi flagged that Amazon had been listing the Leatherman Wave Plus as a 16-tool product when it is actually an 18-tool product. A small detail with real implications for search, content, and conversion. Pi caught it. A human would have had to go looking for it.

Skills, approvals, and the knowledge base

Pi isn't just a dashboard. It executes.

Dave introduced Pi Skills, pre-built automation workflows built around a brand's own standard operating procedures. The demo showed an executive ads summary skill running for Leatherman, pulling brand-specific fonts, colors, and tone from the knowledge base and delivering a board-ready report on demand.

The approval workflow was equally impressive. When Pattern's team flagged new content for the Leatherman Wave Alpha, Pi didn't just show the image and ask for a thumbs up. It surfaced the underlying value propositions the content was built to communicate, the data behind each one, and live performance metrics for the product. Approving content in Pi means understanding why it was created, not just what it looks like.

Brands can also update their knowledge base through natural language. Dave switched a brand color to neon purple live on stage to show just how simple it is. Any change a brand makes is saved and reflected in every future action Pi takes on their behalf.

The content and portal connection

The session also connected Pi to the Portal, Pattern's proprietary AI content capture system introduced in the opening session.

For every product, Pi runs a brand research agent and a product research agent, pulling data from a brand's site, competitor listings, and audience demographics. It then audits existing content and flags whether there is enough reference data to generate AI content or whether the product needs to be shot on the Portal first.

The Leatherman Wave Alpha hit that flag. Every existing asset showed someone holding or demoing the product, which left too little reference data for AI generation to work. Pi caught it, flagged it, and routed it to the Portal automatically.

Where Pi lives

One of the most practical parts of the session was showing where Pi actually lives for brand teams. A Chrome extension brings Pi's intelligence directly into Amazon product pages during live review sessions. Pi is available in the ChatGPT app directory today. Slack integration is live. Claude integration is underway.

The goal is simple: meet brands where they already work. No new tab. No new workflow. Just Pattern's intelligence surfaced in the tools teams use every day.

Pi is available to Pattern brand partners now at pi.pattern.com.

Up next

Dave closed by teasing a deeper dive on agentic commerce coming later at Accelerate 26. Stay tuned for more session recaps throughout the day.

Pi is available to Pattern brand partners today at pi.pattern.com. Prospective partners can learn more at pattern.com/pi.