Page 35 - Harvard Business Review (November-December, 2017)

needing guidance on repairs gets a binder machines that are extremely complex, so How should a company get started
with documentation. We’re developing they can’t be easily automated. Servicing an with AR?
AR on a HoloLens headset that will let the industrial 3-D printer would be an example. First, if you haven’t already done so, you
customer be guided by a remote technician Or work done in semiconductor labs. should design and build your products
who can see what the customer is looking digitally so that you’ll have digital models
at and walk them through a repair. We’re Those are all pretty cutting-edge. Are of them to use in developing AR and VR.
at the early stage. We’ve put together some there less-cool applications that will be Otherwise, you’ll need to create those
prototypes, and we’re sharing those with as important? digital models later, which is complicated.
customers to get their feedback. This doesn’t sound superexciting, but it Second, figure out where AR could generate
In our marine business, we’re working could have a big impact: If people use AR the most value in your operations or
with a coalition of companies on pilot simply to adhere to a best-in-class process, services. I’d gauge that using those three
projects involving autonomous vessels—like it can prevent mistakes and injuries. You dimensions I’ve mentioned: danger,
Google self-driving cars but ships. You can can have the best standard operating remoteness, and complexity of the task. It
imagine starting with small autonomous procedures in the world, but if your workers probably shouldn’t be a priority to add AR
ferries on lakes but eventually scaling up to don’t follow them, it doesn’t matter. AR to a simple machine that’s easily accessible.
container ships. You wouldn’t need large can ensure compliance with processes. For On services, I’d ask where AR could enhance
crews on these ships. If somebody on shore instance, imagine you’re working with an an existing service rather than what new
needs situational awareness of what’s industrial motor and there’s a step in the service you could build from scratch with
happening on a vessel, they could use AR manual that says, “Turn off the power.” AR. It’s much easier to get a customer that’s
technology. We think we could bring this It would be easy to overlook the step and already using some of your maintenance
capability to market within a few years. damage the equipment or get hurt. With services to try an enhanced version. If
AR, the software could say, “Turn off the you and a competitor provide the same
How would that work, remotely power and glance at the switch to confirm service, and yours has an AR component
checking in on an autonomous ship? it’s off.” When you looked at the switch, that allows customers to do some of their
A captain onshore might use AR to see the the AR could take a picture of the state of own work, that creates value for them and
view from the ship’s bridge and contextual the switch, time-stamp it, and record the differentiates you.
information about the ship’s speed and location of the motor using GPS. So you
course and other telemetry data. This is would now be certain that the switch on a How do you see augmented reality and
a case where you’d be integrating virtual motor was off at a specific location and time artificial intelligence coming together?
reality and augmented reality. The VR during a specific step in a process. Today we can create really good artificial
would be the view from the bridge. The AR intelligence that can play Jeopardy! or a
would be live telemetry overlaid on that Do businesses have unrealistic game like Go, but it’s harder for AI to figure
view. If sensors showed that something expectations about AR because of hype out how to respond to situations where
was going on in the engine room, you could on the consumer side? it has no training. It will come up with an
teleport there from the bridge and have a Actually, I think it’s sometimes the informed response, but the outcome can
look around a virtual engine room that had reverse—press about consumer uses of be unpredictable. If you train an automated
AR information superimposed on top of it. new technology negatively influences ship to handle clear skies and a calm sea,
You can imagine needing only a few people the perception of that technology in and a hurricane hits, you don’t know what
actually on board at any time. business. It’s a recurring theme. Think the AI will do. People, at least for quite some
of the press around consumer drones, time, will be better at reasoning in context
What other sorts of jobs do you see for instance, which suggests that they’re in novel situations. So we can imagine that
AR doing? a nuisance or a toy. But of course we’re with the fusion of AI and AR, the AI will
There are three overlapping areas where finding important applications in industry provide a set of recommendations about,
I see AR taking off. The first is in dangerous now for inspecting refineries, pipelines, say, what step to take next in a repair; a
jobs. You want to make sure people have and high-voltage transmission lines. human with the contextual expertise will
the best information possible at exactly Same thing goes for blockchain, which at make the final call; and at that point AR
the right moment, because the cost of first was seen as the technology behind could provide useful guidance. If there’s
not having that—people getting injured, bitcoin, the digital currency used by drug a noise coming from a motor, it could be
equipment being destroyed—is so high. dealers. But businesses are beginning to many things. AI could look at the data and
So I would imagine AR applications in understand that blockchain will have a suggest 10 possible causes and recommend
refineries, chemical plants, construction, huge impact on contracts. AR was seen a few to consider first. But the tech’s
and mining, for example. The second area as a game platform, and there was bad decision about which to follow up on will be
is jobs in remote locations, like on an oil rig press when Google Glass stalled, which based on his experience, his team’s design
or an offshore wind farm, where it’s really may have colored how businesses saw knowledge, what he finds when he opens
valuable to make sure that the people you the technology—maybe as a science up the machine, and so on. He will make
do have on-site have the skills they need. experiment that wasn’t going anywhere. the final call about what the problem most
Third, AR will be really useful in cases But people who actually work with AR in likely is and then select an AR program that
where people are working with products or the industrial space are quite excited. guides the repair. HBR Reprint R1706B

NOVEMBER–DECEMBER 2017 HARVARD BUSINESS REVIEW 61