Methodology

How tools get into AisleVault

AisleVault is independent and vendor-agnostic. We don't take pay-to-list. Here's exactly how a tool ends up in the catalog and how we keep it accurate.

Discovery

Tools enter the catalog two ways: our editors track product launches, funding, and category leaders, and vendors submit themselves through the For Vendors page. Either path goes through the same verification before going live.

Verification

Every listing must have a working website, a real shipping product (not vaporware), and demonstrable AI or robotics capability. We verify claims against documentation, demos, and independent coverage before listing.

Three-axis mapping

Each listing is mapped on the same three axes: industry vertical (where it's used), business function (what job it does), and capability (the AI or robotics capability under the hood) — so buyers can compare like-for-like across the catalog.

Tier assignment

Six tiers across two families. AI: AI-Native (built around AI from day one), AI-Enhanced (existing product with meaningful AI features), AI-Module (an AI feature inside a larger platform). Robotics: Robotics-Native, Robotics-Enhanced, Robotics-Module. Modules always list the AI or robotics product itself, not the parent.

AI vs. Robotics classification

We classify by what the customer actually buys and deploys — not what technology sits under the hood. AI means a software deliverable (models, APIs, apps). Robotics means a physical machine, fleet, or autonomy stack that moves in the world. See the full guide below.

Continuous updates

Pricing changes, new features, repositioned products, sunset tools — the catalog is reviewed continuously. If something is out of date, we want to know.

Editorial review

Before a tool goes live, an editor reviews the entry for accuracy, neutrality, and fit. Featured placements are clearly labeled and never affect organic ranking or matching.

How we classify AI vs. Robotics

A common question: "Wouldn't a self-driving truck company be AI? It's almost entirely AI/ML under the hood."

If "uses AI" made something an AI company, the robotics category would not exist. We classify by what the customer actually buys and deploys — not what technology powers the product.

AI (Software)

The deliverable is software: models, APIs, SaaS apps, or code tools. The customer is licensing or subscribing to a digital product.

  • OpenAI, Anthropic, Cohere — foundation model APIs
  • Cursor, Replit, GitHub Copilot — coding assistants
  • Harvey, EvenUp — legal AI SaaS
  • Perplexity, Mem — consumer AI apps

Robotics (Physical)

The deliverable is a physical machine, fleet, or autonomy stack that moves or acts in the world. The customer is buying hardware capacity or physical automation.

  • Boston Dynamics, Figure — humanoid and mobile robots
  • Symbotic, AutoStore — warehouse automation systems
  • Waabi, Waymo, Aurora — self-driving truck autonomy
  • FANUC, ABB — industrial robotic arms

The edge cases

Autonomy companies (Waabi, Wayve, Aurora, Waymo) are robotics because the customer buys self-driving vehicle capacity — not a software API. The AI is inside the robot (the vehicle), so the product is the robot.

AI Modules inside robots like NVIDIA Isaac, John Deere See & Spray, or Stryker Mako are robotics-module because they exist only inside a robot. You cannot buy them standalone.

Companies doing both get classified by their primary go-to-market. If they sell a robot fleet and also an API, the one that generates the most revenue and customer focus determines the category.

See something off?

If a listing is inaccurate, out of date, or missing, tell us. Corrections are reviewed and applied quickly — accuracy is the whole product.