Visual AI is an image-based technology that automates processes such as filling out punchlists
Construction technology provider OpenSpace introduced a new auto-location service giving customers’ smartphones real-time indoor positioning capabilities on construction sites, whether they’re enclosed or not. The firm’s Visual Intelligence Platform, which includes auto-location functionality, was one of several technologies it announced at its Waypoint customer summit Sept. 9 in San Francisco that uses back-end artificial intelligence.
The reality capture provider’s platform is used today by contractors to document changes on a jobsite via hardhat-mounted cameras. Geolocation via GPS has transformed construction workflows, but a common pain point has been getting it to work indoors once roofs and walls are installed in building projects. Existing solutions to provide indoor location involve additional hardware systems, such as Bluetooth-based beacons.
OpenSpace’s Spatial AI capabilities seek to address this limitation entirely through the software-based AI Autolocation, which compares real-time sensor data from a user’s smartphone with maps generated from pre-existing 360° captures of the indoor space on the OpenSpace platform. As a result, the system can progressively refines its location estimations even as the jobsite changes over time.
«This image-first visual intelligence platform way of working is great because it’s very much like the old way of working [on jobsites], prior to all these forms and web forms.» says Jeevan Kalanithi, OpenSpace CEO. » You look at something, team huddle, move what needs moved and onto next thing.»
He says OpenSpace is re trying to enable existing workflows with AI technology filling in the backend work with features such as Spatial AI.
«That’s what visual intelligence platforms all about. It’s all about putting reality and images at the center of of the work and and moving beyond reality capture,» Kalanithi said.
Suffolk Construction is an enterprise user of OpenSpace’s platform for reality capture and has tested auto-location along with OpenSpace’s voice-to-text AI, VoiceNotes.
«We started a pilot in a project in the southeast that is related to auto-locations and [VoiceNotes],» says Felipe Dominguez, construction technology product manager at Suffolk. «That has been a game changer, when you think about the significant number of time=consuming and the effort that processes like work to complete or punch lists have on our project teams.»
On a typical $100-million project there are usually around 5,000 punch list items and observations over the life cycle, «traditionally, creating those things or those items is time consuming for our superintendents,» he said. «What we’ve seen with this new capability with OpenSpace is that we reduce the time of creating a work-to-complete by around 50%»
Dominguez adds: «Even better is that with the voice-to-field notes function [VoiceNotes], it’s not only saving us time, but we are also improving quality. One of the thing that we see on this project is that around 95% of field notes that we created through these AI capabilities have high quality information.» He notes that «they have location, where they are, where they were taken,» adding that «they have their due date, the trade that they’re responsible for, the assignee and so on. So it’s very powerful for us.»
Dominguez said Suffolk project managers can walk a job and create a work to complete the punch list, but at the same time, they are increasing the quality of what they are capturing.
«We’re very excited about what is coming, and after a really good pilot with them, we’re excited to expand it,» he said.
