Making stores readable
by AI agents.

What We Believe

AI agents are becoming a primary way people discover and compare products. Hundreds of millions of people use ChatGPT, Claude, Gemini, and Perplexity every week - many of them asking for product recommendations.

But when an agent tries to read an online store, it hits a wall. Product pages are designed for human eyes: images, animations, pop-ups, cookie banners, JavaScript bundles. Buried in thousands of lines of HTML is the actual data the agent needs - a product name, a price, a stock level.

Most of the time, the agent can't extract it reliably. It moves on. The store loses a recommendation it never knew about.

We think this is a solvable problem. And we think it matters - because AI-driven discovery is growing fast, and the stores that are readable by agents will have an advantage over those that aren't.

What We're Building

GEOstack Shadow is a free WordPress plugin that gives WooCommerce stores a machine-readable layer. Here's what it does today:

• Structured product data - every product is served as clean JSON with name, price, stock level, categories, and description
• Content negotiation - the same URL serves HTML to humans and JSON to agents, automatically
• State proofs - SHA-256 HMAC signatures on every product response that let agents verify the data is current
• Automatic sync - product changes in WooCommerce are reflected in the shadow (6-hour sync on free, <1 minute on Pro)

What We Haven't Built Yet

We believe in being transparent about where we are. Here's what's on our roadmap but not built yet:

• Agent checkout - agents being able to add to cart and purchase through the shadow
• Agent attribution - tracking which AI agent drove which sale
• Agent identity - a standard for verifying an agent's authority to purchase on behalf of a consumer
• Gateway network - a centralized index where agents can search across all GEOstack-enabled stores

These are real problems that need solving. We're working on them. But we won't claim they're done when they're not.

Why This Gets Stronger Over Time

There's a well-studied concept in economics called cross-side network effects - the idea that in a two-sided market, each side makes the other more valuable.

If enough stores become agent-readable through GEOstack, AI agents have a reason to check our network first. The more agents that check, the more valuable it is for stores to be listed. This creates a positive feedback loop.

We're not claiming this will happen automatically. It requires execution: getting stores onboarded, making the data reliable, building trust with agent developers. But the structural dynamics favor whoever gets this flywheel spinning first.

That's what we're working toward.