A screen grab showing a typical view in Disperse

Technology

The AI start-up working with big contractors

14 May 2021

Felix Neufeld, CEO of AI specialist Disperse, whose digital tools have recently been used by Mace at One Crown Place (https://www.bimplus.co.uk/projects/ai-tech-key-maces-delivery-225m-mixed-use-scheme/), has found construction a great environment for tech start-ups.

“Even the largest construction companies are used to working with small suppliers, so even as a small start-up, you can work with a company like Mace,” he said. The vital thing for Disperse is that it needs access to the data, which in other industries can be a huge headache.

 

We never think we can replace planners or construction managers. Our goal is to supercharge construction managers.– Felix Neufeld, Disperse

Disperse started working with Mace and Canary Wharf Contractors in 2018 and has been growing swiftly since. Today it has more than 120 employees and is working on projects in Europe, Australia, Middle East and North America.

When Neufeld set up the business in 2015, his idea was a tech platform that could work across different industries. He homed in on construction because here he felt he could make the most impact.

Disperse works by comparing digital 360 images, taken at intervals, to previous images and to project information such as drawings and BIM models. Initially humans make the comparisons, with the AI learning as patterns occur over and over. Identifying some tasks can be easily automated; others are more challenging.

“As we get more data and as we are annotating through humans, we automate more and more,” explained Neufeld. Compared to when Disperse started, he reckons that the amount of human input required has fallen by 70%-75%.

With employees spanning all the construction professions, and a particularly hefty team of data specialists to make sense of sometimes conflicting data from multiple sources, Disperse is all about reducing the pain of digital transition and adapting to its customers’ processes.

It can also deliver information as a construction manager requires it – if Excel spreadsheets are your bag, that’s not a problem. “That’s one of the reasons we have been successful,” he said.

Neufeld doesn’t see a fully automated future, more a blend of machine learning and human cognition. “We never think we can replace planners or construction managers. Our goal is to supercharge construction managers and planners so they can focus on the tasks they are really good at,” he concluded.