December 2017
By Sam George

Image: © www.cisco.com

Sam George manages the Switching and Wireless documentation teams at Cisco Systems. He has been exploring Augmented Reality solutions for some time, and is very passionate about using AR to enhance the overall customer product experience. One of his current focus areas is the large-scale adoption of AR at Cisco.

samgeo[at]cisco.com
www.cisco.com


 


 

Operationalizing Augmented Reality in the enterprise

Augmented Reality (AR) is a technology that superimposes virtual information over a physical device in the real world. Consumer deployments in areas like gaming are more common and well known, but how does AR apply to companies that do not necessarily deal with consumer products? And how do you go about implementing an AR solution in a large enterprise?

AR technology has been around in some form or another for several decades now. However, it’s only in the past few years that large enterprises have started realizing its potential as a tool to increase operational efficiency and product usability.

From a usability perspective, some might view AR as the next step in a natural progression: from a command line interface to a graphical user interface to an AR interface, which provides an immersive experience. A recent study by market intelligence firm Tractica estimates that there will be around 70 million active enterprise AR users by 2022. Studies indicate that AR-based instructions significantly improve worker productivity and first-time accuracy.

There are several companies that provide TechCom AR solutions, and many companies are evaluating the use of AR for user instructions. However, not many TechDoc groups in large enterprises have embraced AR today. Let’s find out why.

Current AR technology and limitations

AR investment in the past has been skewed towards wearable design, and the industry has progressed from unwieldy AR headgear to the sleek smart glasses available today. Smartphones with high processing power have also helped make AR a household name. However, there hasn’t been adequate investment or progress made in training AR software to recognize real-life objects. Current AR software recognizes an object mostly based on predefined markers, which can be Quick Response (QR) codes placed on the object, images of a distinctive part of the object, or images of the entire object. While the current technology has helped introduce the benefits of AR to the enterprise, there’s a long way to go before AR reaches its true enterprise potential.

Implementing an enterprise AR solution: a case study

I’ll illustrate the opportunities and challenges of current AR technology with a close look at how Cisco Systems, a worldwide leader in IT and networking, is implementing AR for its hardware products.

A Moonshot

The TechDoc team at Cisco started exploring AR solutions over a year ago. It started more as an experiment to see how we could use an emerging technology to enhance the usability of our core networking products – routers, switches, and wireless devices. Imagine that you’re in a networking lab, and you’re either installing a router or troubleshooting a switch that won’t boot up. You don’t have access to online technical manuals, and you don’t know how to proceed. Now, what if you are able to point your smartphone at the router or switch to get in-context information about the problem you’re facing? You can even see an overlaid animation of the steps to follow to resolve the issue.

Actual reality

However, this scenario is not entirely a reality yet. As mentioned earlier, current AR technology recognizes devices only through markers placed on the device, or by using image recognition to map what you’re seeing to an existing image database. For a large company like Cisco, placing a marker on all its products is not a feasible option, considering the huge existing installed base. Also, image recognition works only when you can predict what your product will look like in its native environment. In a networking lab, a router or switch can be modular, and therefore configured in different ways. This means you cannot use a single image (or even a set of images) to help the AR software recognize the product. There could also be unpredictable cabling around the product, which makes recognition almost impossible.

     Image 1: Device with unpredictable cabling

Image 2: Same device with different configurations

 

Cracking the AR code

Considering these limitations, the TechDoc team at Cisco decided to focus on a single use case that was possible with current technology – AR-enabled instructions for hardware product installation. When a Cisco customer installs a router or a switch fresh out of the box, we know exactly what the device looks like, and there are no cables to disrupt the AR camera’s field of view. An AR solution in this scenario could be a valuable aid to users who are not familiar with the installation procedure, especially if they don’t have easy access to traditional online or printed instructions.

Let’s now take a look at various aspects that you need to consider before rolling out an AR solution in your company.

Canned vs. made to order

You’ve decided to evaluate an AR solution for your hardware products.  Now, should you pick an AR experience off the shelf, or build a customized solution? If you’re looking to conduct a pilot without a significant investment, the obvious answer is an off-the-shelf product. At Cisco, we used Blippar (www.blippar.com) for our pilot. After several cycles of customer feedback and fine-tuning the solution, we’re now scaling it across multiple products. However, there are certain limitations to this approach, the biggest one being that you can’t always customize the available AR functionality in an off-the-shelf product to suit your organization’s evolving needs.

Cisco’s current AR solution provides users with static information about hardware products that is stored in an existing database. For example, if you’re trying to install a Cisco product and it doesn’t work as expected, you can look up a set of static troubleshooting instructions to resolve the issue. However, the AR software does not dynamically change the options available to you based on the success or failure of the installation. We’re exploring our options to build a more intelligent solution that can provide dynamic information about a switch or router that is directly sourced from the network controller (a device that manages all the products in a networking lab). Such a solution will add tremendous value to customers, but it’s obviously not something that you can pick up off the shelf.

The reality of content

AR is a technology that enables user-friendly delivery of content. The content that you display needs to be contextually relevant for the experience to be successful. At a high level, there are two kinds of content you can include:

  • Spatially-aware content, which projects directly onto a part of the device you’re viewing, and is accurate only when it’s projected on that specific part. A good example of this is a label describing a specific part, or an animation that shows how to insert a module into a specific slot.
  • Links to published content that provides information about the product you’re viewing. This is sometimes called "Informed Reality" because it doesn’t augment your reality by overlaying information directly onto the product in front of you but directs you to information about the product (for example, quick start guides or troubleshooting guides).

A good AR experience should have a proper mix of these two types of content. It is critical to tap your internal stakeholders (marketing, customer support) to ensure that you include the most relevant content for your product. At Cisco, the content for every AR experience we create is curated by the customer support and marketing teams.

Target audience and products

So now that you’ve decided what AR solution to use, and what content to include, do you roll it out for all your hardware products? Before you answer that question, you need to understand the level of expertise of your users. Expert technicians installing a device in a lab will most probably not need AR-enabled instructions. However, if your products are handled by low-skilled workers, AR is a great medium to provide instructions.

At Cisco, we partnered with the marketing team to identify high-volume products installed by low-skilled workers. One such example was the Cisco Catalyst Digital Building Series Switch, which is a networking device for smart lighting that is installed when a building’s electrical connections are being set up. AR-based instructions make it easier for electricians with absolutely no networking knowledge to install these switches, thereby eliminating the need for highly-skilled technicians and reducing customer costs.

Scaling the AR solution

I’d now like to touch upon three key tasks involved in implementing AR on a large scale – developing writer skills, creating reusable assets, and driving customer adoption.

The core skill sets for writing procedures remain the same. However, visual thinking is a key skill that will enable writers to develop effective AR-enabled instructions. Writers will also need to learn to work with 3D models if they are animating instructions. The major shift will be that of perspective – reorienting a writer from being a remote instructor to somebody who is virtually present right there in the moment, visually guiding users through complex tasks. At Cisco, we’re reskilling our writers through several in-house training programs focused on visual design and storyboarding.

To provide a consistent user experience across a large set of products, you’ll need to create templates and other user interface artifacts that can be reused across AR projects. You’ll also need to involve your marketing, branding, and UX teams to ensure that the experience is uniform and complies with company guidelines.

A robust customer adoption strategy is key to the success of any innovation. At Cisco, we formed a core team to drive customer adoption by socializing the AR solution with relevant customer segments. The strategy involves developing customer communication plans, adding product inserts, getting adequate exposure on product pages, and tracking overall usage.

Customer feedback

User testing and feedback are critical for an AR solution to succeed. The TechDoc team at Cisco showcased its AR pilot at several customer events, and used the feedback from customers to fine-tune the solution. Most customers felt that AR solutions will make their jobs easier because contextually relevant information will allow them to complete their tasks more efficiently. Customer interactions also helped validate the content choices we’d made in partnership with marketing and customer support. Customers were thrilled to see specific information that they found difficult to locate in traditional user guides.

Effectiveness measures

We’re currently working on defining effectiveness measures that go beyond incidental customer feedback to gauge the overall impact of the AR solution. Off-the-shelf AR software provides inbuilt analytics like the number of users, number of interactions, location of the interactions, etc. A good way of measuring the effectiveness of an AR solution is to analyze the trend in customer support cases that relate to the specific areas targeted by the AR experience. Establishing a downward trend in support cases can also go a long way towards convincing management to invest in a customized solution that further enhances the customer experience.

Is true enterprise AR a distant reality?

As discussed earlier in this paper, the true potential of AR in the enterprise will be unlocked only when AR technology evolves to a level where it can recognize modular objects in unpredictable environments. The AR industry is currently exploring Artificial Intelligence and machine-learning algorithms to help AR software recognize a hardware product even if it looks very different across multiple customer installations. Given the current state of research, I believe it’s going to take the industry at least a year or two to develop AR technology that can achieve accurate object recognition without markers.

Enterprise AR in its current avatar offers a lot of opportunities for companies that are looking to augment their documentation offerings and enhance the product user experience. AR solutions are relatively well developed in certain industries like aviation where precision is of the utmost importance. For others, I wouldn’t go to the extent of calling AR an indispensable tool. However, in my experience, an AR-based technical communication solution, when targeted at the right customer base, helps increase first-time accuracy and worker efficiency. If your documentation reality is still rooted in HTML and PDFs, I would definitely recommend that you augment it with a judicious dose of AR!