Inside an Amazon Fulfillment Center: an Immersive Twin with AWS & Matterport

Adapted from the AWS video on the AWS × Matterport × Treedis collaboration, featuring Gabriele Biagini, Solutions Architect at AWS.

A fulfillment center is hard to show the world.

Amazon fulfillment centers are vast, highly automated environments, packing stations, conveyors, and fleets of mobile robots moving inventory around the clock. Public tours have run since 2015 as part of Amazon’s transparency policy, but only a limited number of people can ever walk the floor in person, and a scheduled visit can only show so much of how the operation actually runs.

The goal was to let anyone, anywhere step inside a working fulfillment center and understand it, the packing stations, the robotics, and the flow of an order, without booking a visit.

Capture with Matterport, bring it to life with Treedis, add AI with AWS.

The space was captured with Matterport into a dimensionally accurate 3D digital twin. Treedis’s no-code platform then turned that capture into an interactive, guided virtual tour: each key area and asset, the outbound packing stations, the Pegasus and Proteus robots, and the equipment that services the floor, is tagged in spatial context with information, media, and step-by-step workflows.

On top of the experience, the team used Amazon Bedrock to power an AI assistant, so visitors can ask questions about what they’re looking at and get grounded answers, all delivered through a 3D hosting platform on AWS.

• Matterport captures the fulfillment center as a true-to-life 3D twin.
• Treedis adds the working layer, no-code, with tagged assets, guided journeys, and the virtual tour itself.
• Amazon Bedrock powers the conversational AI that answers questions inside the twin.

A personalized journey through the floor.

Rather than a single fixed walkthrough, the twin delivers a personalized journey. Visitors move through the fulfillment center at their own pace, opening guided steps on individual stations and robots, watching embedded media, and asking the AI assistant how scheduling, robotics, and each part of the operation fit together.

The same twin doubles as an operational tool, with visualization for predictive maintenance that surfaces the condition and context of equipment directly on the assets where it matters.

A repeatable pattern for spatial experiences on AWS.

The fulfillment center twin shows a repeatable pattern: capture a real space once with Matterport, build the interactive and operational layer on Treedis with no code, and add grounded AI with Amazon Bedrock on AWS. The outcome is one digital twin that is both an experience open to the world and a working layer for the people who run the facility.

Adapted from the AWS × Matterport × Treedis video featuring Gabriele Biagini, Solutions Architect at AWS.