(This column originally appeared in Forbes)
AI applications are slowly but surely being used by businesses and their employees to write better emails, research topics, supplement ideas, create policies, review contracts and to perform basic agentic tasks. But think about what an employee who’s not a desk — a project manager, construction supervisor, safety engineer — needs in order to do their job. That’s where companies like OpenSpace are jumping in.
Visual AI
OpenSpace helps its customers — mainly in the construction industry — to capture what’s going on in the field and then time-stamp it as a visual record so that managers can track the progress of a project, identify issues and get a better handle on their timetables.
The 350-person company does this by capturing what’s going on using 360° cameras carried by a person walking the site (their original approach), smartphones (for certain workflows, especially field notes) and now drones (added more recently) to create an output that is a navigable, visual layer with a Google Street View look-and-feel to it so that project managers can move through a project like a map using different timestamps from the past.
Imagine a 24/7 observer reporting progress back to the office at regular intervals in a way that gets immediately translated into data and integrated with a project management like Autodesk or Procore or a building information modelling (BIM) system. Or using this as an AI-driven progress tracking that mirrors what an experienced project manager would conclude by walking the site.
AI’s Next Wave
Jeevan Kalanithi, the CEO of OpenSpace believes this is the next wave of AI. It won’t be text based. It’ll be visual and spatial. And that’s what his company is doing, to the tune of almost 20 million images captured by their customers every week.
“A lot AI today is very oriented on text and documents,” he said. “But the real world is not text, it’s real physical reality. AI systems like ours will digest not just a document or spreadsheet, but an image of the real physical world.”
People can interpret reality in any number of different ways. But a picture’s worth a thousand words. And it’s hard to argue an interpretation when it’s right there in front of your eyes. AI is now being used to do that — and, most importantly, help to project where things are going. Is the project on track and on time? Are there any issues on the horizon?
This is particularly important for companies in the construction world, where there is no “undo” button once work is completed. OpenSpace uses AI to create a persistent visual record over time, according to Kalanithi, it’s the “source of truth” and one that can be revisited, audited and analyzed.
AI As Infrastructure
Platforms like OpenSpace usually don’t exist on their own. They’re a form of middleware, capturing video and image information from the field, using AI to interpret the data, and then passing it through to other applications. Kalanithi is clear that OpenSpace is not trying to do everything.
“We want the data to flow to where decisions are getting made,” he said. “This is infrastructure, not interface.”
Kalanithi agrees that AI has yet to move the needle for real-world workers. But he believes that this will change in the next three to five years as the technology is able to more reliably interpret real-world data and apply it to internal systems and processes.
“The next breakthrough phase of AI won’t be felt in offices,” he said. “It will be felt on job sites, factory floors, warehouses, and physical environments. That’s because people running organizations want the facts of the job, not just opinions.”
