The Challenge: Manual documentation of thousands of products
For Betonmast, a leading construction contractor, the new BREEAM standard (version 6.0) introduced in 2022 created a significant hurdle. The standard set higher requirements for documenting and declaring the content of all products used in a building, particularly concerning chemical content.
“We can’t manually read documentation on thousands of products per project. It’s nearly impossible. Timewise, this is simply not feasible.”
— Siv Tallang-Vold, Environmental Manager at Betonmast
Betonmast had more than 5,000 documents without a standardized format, and their existing system couldn’t assess the content. This meant human eyes had to manually review every single product documentation to check for over 200 environmentally harmful substances listed in the European REACH regulation.
A complex, error-prone task
The task wasn’t just time-consuming; it required specific expertise to interpret Environmental Product Declarations (EPDs). These documents lack a standard format, coming in various layouts with text, tables, and markers.
“First, you have to determine if information about the chemicals is even included in the documentation. Then, it has to be correctly interpreted and registered,” explains Tallang-Vold, noting the high risk of human error in this manual process.
The Solution: AI-powered analysis
Betonmast partnered with Findable to tackle this “impossible” challenge. Findable’s AI technology was deployed to automatically read, analyse, and categorise the thousands of unstandardised documents.
“You can imagine searching for ‘REACH’ in a PDF and trying to find the answer. That works fairly well. But what do you do when there are 10, 100, or 5,000 files?” says Lars Aurdal, Co-Founder and Head of Research at Findable.
The Results: A week’s work in 10 minutes
The impact was immediate and dramatic. What was previously an insurmountable manual task became a quick, automated process.
“We were able to offer Betonmast a week’s worth of work in about 10 minutes.”
— Lars Aurdal, Co-Founder at Findable
Findable successfully automated the detection of chemical information across thousands of disparate documents—a task that previously required manual review by experts.
“With Findable, we’ve managed to crack this challenge in record time. I’m truly impressed with the results.” — Siv Tallang-Vold