I think you’re right that the demand for a huge project like the one in the article is going to be extremely limited, but I can imagine a lot of bigger companies finding it interesting enough to dabble in. There are many people still out there who haven’t been exposed to this new type of interactive walkthrough, and it’s still an exciting idea that businesses may yet be exploring.
It definitely takes a long time to create an entire product and workflow from scratch, but I think using a game engine could be much faster in the long run if you are comparing it to traditional rendering (while factoring in the interactivity). Once you have a library of furniture and materials, it becomes easy to throw together a scene into the pre-programmed game you have. Modeling the unique items for each location is the biggest problem, but it could only take a couple days depending on your team. There is almost no rendering cost, and iteration time is impressive.
I feel like Unreal’s realism is mostly limited by the user. Obviously there are technical limitations that can’t be overcome, but I’ve seen some jaw-dropping screenshots from realistic ArchViz projects posted in the forums.
My company is Theia Interactive. We’ve been keeping in touch with some interested real estate owners, and we just launched a campaign of sorts to get dedicated buyers.