Reducing visual documentation time by 90% using AI on AWS with PlanRadar

With over a decade of experience providing software solutions for the construction and real estate industries, PlanRadar understands that documentation and communication can be a persistent pain point for project teams. Traditional photo documentation requires thousands of manual photos and hours of work, often resulting in gaps or misaligned images.

“In construction, there’s always a shortage of labor,” says Constantin Köck, chief technology officer and cofounder of PlanRadar. “Everything that saves work hours and is easy-to-use is an advantage.”

Building on its prior use of AWS, which dates to 2013, PlanRadar sought to create a low-effort, General Data Protection Regulation (GDPR)–compliant visual documentation solution that would integrate with its existing solution. The company’s container-based property technology platform already runs on Amazon Elastic Kubernetes Service (Amazon EKS), which streamlines Kubernetes operations by automating cluster infrastructure management. This foundation provided the scalability and reliability PlanRadar needed to expand into new AI-powered capabilities.

To begin a SiteView session, the customer simply attaches a 360-degree Insta360 camera to their safety helmet and walks the site. “Instead of taking thousands of photographs manually, you just turn on the camera and walk,” says Clemens Hammerl, chief product officer and cofounder of PlanRadar. “Our product does virtually all the work.”

When the walk is complete, the recorded video is uploaded to Amazon Simple Storage Service (Amazon S3), an object storage service offering industry-leading scalability. “We chose Amazon S3 for storing our video data because it offers affordable, reliable storage,” says Köck. SiteView extracts metadata from the footage, combines the front and rear camera views, scales down the video, and blurs any people captured to maintain GDPR compliance.

“It’s inspiring to see how PlanRadar uses our 360 cameras to power AI-driven construction workflows,” says Jason Jia, chief technology officer of Insta360. “By combining 360-degree imaging with AI, they’re not only transforming how sites are documented but also generating actionable insights that deliver real value.”

SiteView then uses machine learning algorithms to calculate distances and analyze floor plans. For these AI workloads, PlanRadar uses Amazon Nova Foundation Models, which deliver frontier intelligence and exceptional price performance, in Amazon Bedrock, a comprehensive, secure, and flexible service for building generative AI applications and agents.

Once processing is complete, SiteView extracts the key frames from the footage and builds an interactive walk-through of the construction site. The solution recognizes previous recording paths from images captured on different dates and aligns new key frames to the same spatial coordinates. This empowers users to compare progress over time—essentially offering a time-machine view of their project. The result is an experience where users can tour the construction site in space and in time, letting teams check what lies behind walls or ceilings without additional invasive inspections.

PlanRadar’s customers now benefit from an AI-enhanced, intuitive, and cost-effective documentation process. “There’s no extensive training required, unlike with some other camera systems,” says Hammerl. The average investment for customers is the 500-euro cost of the 360-degree camera.

Customers can document 500 square meters of a construction site in approximately 5 minutes with SiteView, resulting in a 90 percent reduction in time compared with manual photography. SiteView delivers positioning accuracy of 97 percent, which exceeds industry alternatives. And because the solution is directly integrated with PlanRadar’s software, users can manage all site documentation in one place, greatly improving the user experience. “We automate virtually everything for customers,” says Hammerl. “Our software and AWS do the heavy lifting; the user just needs to press start.”

Looking ahead, PlanRadar will continue to incorporate more AI features into its solutions using foundation models in Amazon Bedrock. With over 100 foundation models available, Amazon Bedrock provides a vast selection of models that PlanRadar can choose from for future use cases—including automated site safety inspections. The company also plans to adopt Amazon S3 Vectors—cost-optimized, AI-ready storage with built-in support for storing and querying vectors at scale.

Since going live with SiteView, Plan Radar has had approximately 50 new access requests per month. That surge in interest reflects the benefits that customers are experiencing: dramatically reduced documentation time, higher accuracy, and better ways of working.