Visual Customer Engagement: Glancing Beyond Co-BrowsingVisual Customer Engagement: Glancing Beyond Co-Browsing

The ability for agents to co-browse the desktop Web pages or mobile app screens customers are viewing has created a significant business, with some estimates suggesting this market is approaching $1 billion in sales annually. Although co-browsing has been available for a long time, there’s a demonstrably better way of enabling co-browsing in the contact center.

Last month at Enterprise Connect 2019, I met up with Glance Networks, a company I first encountered nearly 15 years ago. At that time, Glance billed itself as a simple, inexpensive “see what I see” alternative to Cisco Webex, LogMeIn’s GoToMeeting, and the like. At Enterprise Connect, Glance convinced me that it had transformed itself into a visual customer engagement solution worth noting.

Three Competing Ways to Co-Browse

Most of us have used some type of co-browsing for “seeing what I see” at one time or another. This technology is really at the heart of many of the desktop collaboration and Web conferencing tools we use to share screens. For customer support, co-browsing can be extremely useful because it allows an agent to see what the customer is viewing and provide guidance on what to do and how to do it. I’ve personally used a number of these co-browsing solutions for a manufacturing business in which I have interest.

There are basically three ways to do screen sharing, which I briefly describe below.

Generating Screenshot Representations of Web pages

One obvious way to co-browse is to send a snapshot of the Web page or mobile screen across the network using an application both parties run. Under the hood, this application uses the page’s document object model, or DOM (the HTML and CSS feeds), to create a representation of the page, as opposed to taking a real screenshot. Because the application doesn’t really render the page, the screenshot may not be 100% accurate as to the page’s actual display. The application will transmit the image at some predetermined frequency to imitate movement or animation on the Web page, a capability that you could refer to as Lo-Fi video.

This method is simple, but it has some drawbacks.

First is that such an image would typically be compressed before being sent, and that process would affect viewing quality on the far end. For example, expanding the image would likely cause pixilation that reduces image fidelity.

Second, Web application security models require all graphics on a rendered Web page to reside in the same origin, which is defined as a combination of URI scheme, host name, and port number. If they aren’t in the same origin, they won’t be readable. This policy prevents a malicious script on one page from obtaining access to sensitive data on another page through that page's DOM. This means that graphics from multiple sources outside of a page’s origin won’t render.

Third, annotations don’t work well in this scenario: If an agent circles a place on the screen or draws an arrow, these markups often won’t display very well.

Click below to continue to Page 2

Sending a Video Representation

Another way to co-browse is for an application to grab a screenshot, or some defined portion thereof, at some specified rate, and convert these images into a compressed video stream. In such a scenario, the application typically only sends the changes in the screen images between screen grabs. This is how Webex, Microsoft Skype for Business, and many other tools like them work. Use of these tools often requires a license or subscription. WebRTC, a free tool built into some browsers, is an alternative. However, it too has some challenges when used for co-browsing.

For example, screen sharing, via WebRTC’s data channel, can hog resources. The device on which WebRTC is running sends video at 30 frames per second, with a separate stream sent to each participant. So, screen sharing via WebRTC can consume a computer’s CPU resources. While contact center screen sharing will typically only be between two parties, the agent and the customer, the bandwidth for sending WebRTC-based HD video begins to become onerous for contact centers with hundreds or thousands of agents.

In addition, WebRTC still doesn’t work seamlessly across all Web browsers. While all major browser vendors have announced WebRTC support, you still regularly come across disclaimers like, “works with FireFox version xx and above” or “works only with Chrome version xx and above.” Microsoft Edge has some Object Real-time Communications (ORTC) capabilities that sort of work with WebRTC, but last I checked the Edge browser’s data channel didn’t work.

Rendering Using the DOM Elements

A third way to share a Web page is to use an application that sends the actual DOM elements (HTML, CSS scripts, etc.) from the customer’s Web page to the agent for full-fidelity viewing. This is the most accurate way of sharing a Web page or app screen with an agent because the fonts remain crisp and the images maintain their quality, with 100% compatibility across any browser in use. If an agent marks up the screen, the annotations get inserted into the HTML code and are displayed perfectly on the customer’s side.

While this sounds simple and rather elegant, the reality is that screen sharing using the DOM elements is really difficult to do well. The reason is that any screen sharing tool that uses this method must accommodate everything the browser does; otherwise, the pages may not display correctly. Furthermore, the screen sharing solution needs to understand all HTML versions because the agent cannot determine which browser and version of that browser a customer is using.

Screen sharing using the DOM method has some significant advantages:

However, DOM-based screen sharing can have a significant downside: the use of a proxy server. To simplify the process, some screen sharing using DOM methodology require a proxy server to render the shared DOM elements. However, a proxy server is going to be bad for many customer service organizations because:

Click below to continue to Page 3

Back in a Glance

As mentioned earlier, I’ve known Glance Networks for many years, but I had sort of discounted the company as a poor man’s Webex. While super easy to use and install, Glance was intentionally limited in functionality. It was a great “see what I see” tool, and that was pretty much all early versions of Glance did.

At Enterprise Connect, however, I learned about Glance’s transformation into a powerful purpose-built customer engagement platform for organizations that need secure co-browsing between agents and their customers. I was surprised to discover that some of the largest banks and financial services organizations, insurance companies, healthcare organizations, retailers, manufacturers, travel companies, and airlines have already installed the platform.

The Glance co-browsing technology is enabled by adding a few lines of Java code to a Web page. Customers using Glance usually instrument every customer-facing page on their sites with this code, which when combined with customer journey solutions, allows the agent to meet the customer where he or she is on the site. In addition, Glance has a mobile API enabling co-browsing in Apple iOS and Android apps. Because Glance uses the DOM elements, an end user doesn’t have to download any software, and the solution works on every kind of Web browser.

Glance has invested deeply in understanding how browsers work at the core, and it uses the DOM element rendering model for screen sharing. Because of this deep knowhow, Glance’s co-browsing technology doesn’t require a proxy server. Co-browsing sessions are point to point, and nothing is stored on Glance’s servers. Consequently, Glance’s solution can easily pass an enterprise information security review.

Keeping Sensitive Information Private

With Glance, organizations can use HTML or CSS tags to keep sensitive information from displaying when co-browsing. For example, if a customer is logged onto a company’s Web page or is using an app that displays personally identifiable information such as a Social Security number, credit card data, or private medical information, such fields would be tagged as private and redacted on the agent’s screen. With Glance’s support for tagging, contact centers can remain compliant with GDPR, HIPAA, and PCI regulations.

In addition, Glance-enabled contact center agents can only see those Web pages or mobile phone apps that have been instrumented with Glance’s technology. Consequently, the agent doesn’t see background images or any other windows or app screens that a customer may have displayed on his or her PC or mobile device.

The No-Frills Solution

As it did in its early days, Glance provides a powerful co-browsing capability that works well without frills. For example, the platform doesn’t support voice, nor does it offer recording, since Glance figures that contact centers already have such capabilities. If a contact center wants to record the agent’s screen for audit or regulatory purposes, it can use the same screen recording application it already has in place. Glance doesn’t store any customer data on its servers.

Glance has enabled the display of agent video on the user’s screen for both Web pages and mobile apps, per customer requests, for increasing caller satisfaction and reducing caller agitation. The opportunity to view agents not only creates a stronger bond between a contact center and its customers, but also leads to less caller rudeness and vulgarity, Glance customers have shared.

Results, SLA, and Cost

Some Glance customers have reported stunning results, such as 10% higher first call resolution rates and an 18% increase in transactional Net Promoter Score, thus demonstrating the economic power of visual customer engagement.

Glance has a small reseller channel, plus it has some OEM partners that embed it within their own solutions. Glance integrates with solutions from ServiceNow, Salesforce, Microsoft Dynamics, Zendesk, and Moxie. The company offers a standard availability service-level agreement of 99.9%, which can be increased to 99.999% via a broader support relationship. Expect to pay between $700 to $1,000 annually for a Glance license, which is a shared license that can be used by multiple agents around the clock. The company also has utility pricing for organizations that deploy Glance beyond the contact center.

Disclaimer: The author has not been compensated by Glance Networks for writing this article. He believes this is an example of innovative contact center technology that would be of general interest to the communications market.