2017 has been a pretty deflating year for virtual reality.

In 2016 we were promised that the era of consumer VR was finally arriving, and 2017 was supposed to be its shining moment. Facebook had really big plans for Oculus. The New York Times doubled down on giving away Google Cardboard to its subscribers. Valve’s clever Lighthouse offered a solution for tracking. Playstation VR launched with great promise.

Personally I’ve always been more bullish on AR than VR for mainstream adoption, but I worked on several VR projects in 2016 and there was a palpable feeling of excitement. The technology had improved very quickly and was finally quite good. Very talented creators were experimenting with the medium, resulting in some really interesting experiences. And all of that is still true.

But all the momentum really seemed to come to a screeching halt around the end of the year. Holiday sales of the various headsets had been lackluster at best. At CES in January there was a general consensus of disappointment at what the VR industry had to show. Everyone was reminded that there are still big headwinds for consumer adoption of VR. And the hype cycle hasn’t recovered yet.

But wait, didn’t we skip a step? Sure, virtual reality is still too cumbersome and expensive for most consumers. But where VR can facilitate commerce, clearly the ROI comes out positive in a large number of use cases. We should still expect to see more widespread adoption by the business community, even if consumer adoption is lagging. Actually most of the VR companies I worked with in 2015 and 2016 were working to solve business problems more than to extend entertainment. One created virtual walk-throughs to help sell real estate. Another let college students take virtual tours of campuses.

Probably the industry or industries with the most clear need for virtual reality solutions are architecture, engineering and construction, collectively known as AEC. Trimble, one of my former clients, serviced this market heavily. In fact they made a version of SketchUp for the Hololens, an HMD that’s technically AR rather than VR, but the concept is more or less the same. Actually the AEC market might be one of the few areas I can think of besides entertainment where VR makes more sense than AR.

Back when I was in college I studied engineering and architecture, and I can recall being aware that an inherent ability to visualize physical spaces was pretty much a pre-requisite for an architect or a structural engineer. It was just a part of their professional skillset. This was back in the days before CAD, when we all had drafting tables and sketched with pencils. Of course that spatial skill is still a quality most architects have, but it’s less of a requirement now. Because VR can now allow anyone to virtually inhabit a physical space that hasn’t been built yet. Not-built-yet being a common state of projects for most the AEC realm.

And of course now these fields are completely computerized, beyond CAD to detailed BIM files that inform the entire build life cycle. BIM serves a completely different purpose and therefore differs greatly from the 3D modeling techniques used for VR, which have been largely inherited from gaming. But it’s still a digital 3D model and there’s enough overlap to give a big head start to a VR experience by starting with a BIM file.

In fact, this is such a natural fit that the tools for bringing BIM into VR are improving rapidly. Autodesk now even has its own game engine, called Stingray, specifically designed to make real-time renders of BIM files created in its flagship BIM product Revit. Unity and Unreal, the two industry standard game engines for VR, can also make relatively quick work of converting BIM from Revit or Rhino into a working VR environment.

The resulting experiences from straight conversion are relatively rough. The refinement of textures and lighting that lend realism to game design are not available from BIM and must be redone (or done in the first place) within the game engine, if such realism is the goal. But is that the goal? I’d expect that in many cases the goal is simply to understand the spatial qualities of the design. Even if architects are better than most at imagining these spaces, no one can expect their clients to be able to do so. So placing clients virtually into a proposed design must surely be a game changer for the architecture industry.

Of course some clients will want that higher degree of realism. And it helps if they’re on the cutting edge of 3D rendering. Architecture giant Gensler evidently used NVIDIA’s own iray real-time render engine to visualize in VR the design of their Silicon Valley headquarters. That seems like perhaps the ultimate synergy for business use of VR.

But other big architecture firms seem to agree. HOK wrote about their use of VR as a design tool. Rockwell Group has a lab that’s experimented with VR. Wired spoke to SHoP, Ennead and SOM for an article late last year.

So despite the overall disappointment at the virtual reality industry from the technology press, I think it’s safe to say we can still be excited about the bright future for VR. As long as we know where to look. And it’s not necessarily dependent on more headsets under the Christmas tree.