No Jitter is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

WebRTC and the Contact Center of the Future

Like many people, I avoid making calls into contact centers. I hate navigating through interactive voice response (IVR) systems, entering or speaking long account numbers, and then waiting, before hopefully finding a good agent who can quickly help me with my problem. It often seems much easier to go to the company's website and do some quick searching through its self-help system. In the event that I need a live person, I usually try a Web chat session.

Much of this is by company design; agents and the infrastructure around them cost a lot of money, and Web-oriented self-service is much cheaper, and multi-tasking Web-chatting is much more efficient. But is this really the best way to handle every customer interaction? Is the company really maximizing the value it can get out of me and its own profitability? There must be a better way, for at least some calls.

Fortunately a better way is coming. The Web, with its newest feature, WebRTC, is poised to change the way we think about contact centers.

Customer perspective
Let's first think about the ideal contact center situation from the customer's point of view. Preferences for obtaining information from a contact center or website will vary considerably based on the situation, and may even change during an individual transaction.

For example, if I am usually fine checking my balances online, there is no good reason a person should be involved for a simple information transaction like that. However, if I see that hundreds of dollars have mysteriously been withdrawn from an ATM on the other side of the globe, then I am going to want to talk to someone that instant. Even better would be for an agent to automatically request a session with me due to suspected fraudulent activity. Ideally, I would also be able to see that agent via video in a situation like this. After all, I have just been digitally robbed, and I want the extra empathy and authenticity of seeing someone live helping with my situation.

If I can access my account information from any device I happen to be in front of, shouldn't I also be able to click a button to talk to a live agent anywhere, too? Some other desirable improvements:

* How often do we hear, "This call may be recorded"? Come to think of it, I would also like access to the recording and full text transcript of the interaction for my records.

* Obviously, I want my problem solved as quickly and effortlessly as possible since I did not want the problem in the first place. If my problem requires three experts to all be on the call simultaneously to quickly sort it out, then that's what I want.

Company perspective
Now let's look at it from the contact center perspective. The company wants its customers to be delighted, but it cannot afford a personal concierge for everyone. The trick is inserting the live agent when the customer really needs the agent, and where it will lead to a profitable outcome for both parties. After that, the critical metric is resolving the problem as quickly as possible, both to save the customer's time and increase agent efficiency.

Using analytics and common sense, the company should be able to identify when a customer is starting to have issues via the Web interface that would be more effectively resolved by a live agent. Depending on the situation, this might mean the agent comes in and out of the transaction as the customer looks through the website. More profitable customers are going to get more attention. Fraud is also a huge expense, and the fear of fraud causes customers to spend less. Real-time voice and video biometrics to authenticate users and identify suspicious behavior would be a huge benefit.

Once someone is on the phone, the call quality had better be superb--nothing is worse for the customer or agent than wasting time repeating everything. In addition, some problems are not easy for the customer to explain. If a picture is worth 1,000 words and a customer call is eight minutes at an average of 120 spoken words per minute, you don't have to do the math to know it would be enormously helpful if the customer could send a picture or add a live video stream if the situation warranted it to reduce agent time.

The contact center manager has even more issues about which to worry. First, he needs to have a sufficient pool of staff to answer customer queries. Agents aren't just locked up in a call center somewhere; increasingly they are in remote places, and they need to connect to contact center systems no matter their location. In addition, the contact center needs to maintain its regulatory compliance policies--which often means recording. They also need to allow supervisors to coach in whisper mode or be added to the call as needed.

Gathering as much data as possible for Big-Data analytics is critical for fine-tuning and improving the entire process. The text transcript of the interaction, sentiment analysis of the audio--potentially even video if that was involved--is critical for analytics and improving the customer experience. This data could also be useful to the agents in real time, as they are handling multiple customers via several different channels simultaneously. This information would help them, and their managers, keep track of everything and focus on resolving dissatisfaction.

WebRTC is how we get there
All these technologies exist today. We see some of them often, such as speech recognition and recording. Some, such as audio and video-based sentiment analysis, are nascent, but coming. However, we rarely see many of these technologies used all at the same time. Why not? One major reason is the inability to organize and process all these disparate services in today's contact center environments. It is just not practical to do. This is where WebRTC comes in.

Unlike most of the Web, which has static, long-lived, largely text-based content, telephony is based on real-time, ephemeral streams. Passing these streams around in the TDM and traditional VoIP domains is an expensive and complicated effort. The Web world provided no good way to do this either.

This is where WebRTC changes things by providing a low-cost and practical transportation mechanism to send these streams through the Web. In fact, WebRTC's open source media engine library is being adopted by many traditional VoIP companies because of its quality and free community support (partly due to Google's investment). Use of this engine helps to minimize development costs and prevent lower-level interoperability issues.

In the Web world, it is not uncommon to invoke dozens of services simultaneously in sub-second speeds as needed. For example, a simple Google Search touches more than 200 services in Google's cloud. Web services can do this because they have flexible, interoperable, yet robust interfaces. WebRTC, with its standard Javascript APIs, brings these same kinds of interfaces to the telephony world. These APIs fit in nicely with the other Web services that have been out there for years and are needed to make the above happen. Because WebRTC is designed for the Web, software that uses it can leverage existing infrastructure, architectures and investments that already exist to power the hundreds of millions of websites and thousands of Web services available today.

Yes, there are other ways to reach this contact center nirvana, yet none of them have been able to achieve it. Real-time synchronization of many disparate, computationally intensive media processes is certainly not easy, but leveraging the Web approach to this problem makes it a lot more viable. WebRTC is just getting started, but it brings with it the right mix of technology and community enthusiasm to make this future a reality.

Chad Hart is senior director of product marketing for Dialogic where he is responsible for strategy, messaging, and ecosystem development for Dialogic's PowerMedia media processing and WebRTC portfolio. Chad was previously at Oracle/Acme Packet where he most recently focused on WebRTC, OTT communications, and network function virtualization (NFV). His 14 years of experience in the communications industry includes business intelligence, product marketing, product management, analyst relations and as a syndicated industry analyst. Chad is also chief editor of webrtcHacks.com, a blog for WebRTC developers.