What is an IVR?

Spoken | August 9, 2016

Part of our ongoing call center technology glossary series: what is an IVR?

Welcome to the second installment of our call center technology glossary series. In this installment, we’re going to be looking at the acronym IVR. If you’ve ever called a customer service center and been greeted by an automated recording rather than a live person, then you are familiar with an IVR system.

What is an IVR?

What does IVR stand for? Interactive Voice Response. The IVR works closely with the Automated Call Distributor. It is the software that allows customers to interact with the company via the voice channel. IVRs often serve three main functions:

  1. Improve the customer experience A primary function of the IVR is to route the caller to the correct queue. Billing queries go to the billing department, and sales calls go to the sales department, for example. With misrouted calls and transfers being a key indicator of customer dissatisfaction, getting to the right queue the first time is a key indicator of call quality.
  2. Service high call volume During periods of peak call volume, an IVR can be a useful tool. First, the IVR can offer self-service for some types of inquiries, reducing the agent load. Second, the IVR can give an estimated wait time to allow callers the option of a callback (virtual hold) or of IVR or web self-service. Third, the IVR can help to identify callers quickly, so that the agent already has the caller information when he does answer, thereby shortening the call length.
  3. 24/7 service While it may be cost-prohibitive for many organizations to staff phone lines outside of regular business hours, an IVR can take messages or provide automated information so that the caller might be able to self-service. For example, if a caller simply wants to know shen her package was shipped, the shipping information could easily be delivered through the IVR, even at 2:00 in the morning, and the caller would never need to speak to an agent nor call back during business hours.
  4. Reduce cost It would be disingenuous not to acknowledge that IVR minutes cost far less than agent minutes. One of Spoken’s customers was able to reduce the cost per call by 15% through effective IVR automation.

Download the case study: 15% reduction in cost per call with the Spoken Smart IVR

A brief history of speech recognition

DTMFpad.jpgIn this age of Siri and Smart IVR, it’s hard to remember when speech recognition systems were more rudimentary, shall we say. Prior to 1963, telephone numbers were dialed by users with a loop-disconnect (LD) signaling using rotary dials. DTMF, or dual-tone multifrequency signaling, was first developed by Bell In 1963; it used a voice frequency band to deliver information using tones.This revolution spelled the death of the dial phone and the birth of “for sales, press one.” Under the trademark brand Touch Tone, land line customers could now purchase a phone with buttons that were pushed rather than dialed.

darpa-680x346.pngSpeech recognition research was going full swing in the 1960s. At that time, speech recognition followed the trigram model, by which each word required a 350 millisecond pause afterwards in order to be recognized. So speech recognition sounded something like this: “I… want… sales.” Then, in 1969, John Pierce at Bell System famously defunded the speech recognition research that Bell was doing at the time. He compared speech recognition research to “turning water into gasoline, extracting gold from the sea, curing cancer or going to the moon.” Happily, the research continued despite Pierce’s beliefs. In fact, in 1971, DARPA, the Defense Advanced Research Projects Agency, part of the Department of Defense responsible for emerging technologies for use by the military, established the Speech Understanding Research (SUR) program. The SUR program had a goal of developing machines that would understanding continuous speech, without that 350-millisecond pause after each word.The project was somewhat successful, but it took computers 100 minutes to decode each 30 seconds of speech!

In the 1980s and 990s, huge strides in speech recognition were made. Under Fred Jelinek’s lead, IBM created a voice-activated typewriter, known as Tangora, that had a 20,000 word vocabulary in 1985. In the 1990s, speech recognition vocabulary actually exceeded average human vocabulary and became speaker-independent.

In the 1990s, speech recognition technologies became commercially available, and call center technologists took notice. Companies began to invest in Computer-Telephony Integration systems, which connected the analog voice systems with the digital customer relationship management databases. This meant that call centers could use speaker-independent speech recognition instead of DTMF signaling. Now, callers could say “sales” instead of pushing one on the touch pad!

In the following years, voice and speech recogntiion technologies became less expensive and easier to deploy, thanks to ever-increasing CPU power and the introduction of the VXML standard. With those innovations, IVR became a cost-effective call center solution.

An IVR illustration

You’re likely familiar with the drill. You make a call to a particular business, you are asked to state the purpose of your call by answering a question or by pressing a certain number. Once you answer or press said number, you are routed to the correct department. The software that does this is the Interactive Voice Response. The infographic below illustrates how this works.










Spoken’s approach to the IVR

All that being said, let’s acknowledge that many IVRs can provide a frustrating experience, usually due to thoughtless or inexpert design. A well-designed IVR call flow is a thing of beauty: the caller always feels understood; the call is always routed to the right queue the first time, and the caller is accurately identitied within the CRM system so that the agent can have a successful interaction with the caller.

However, we all know that there are many roadblocks to that IVR utopia:

  • Unclear utterances A dog barking or a siten in the background will break most speech recognition IVRs and ask callers to repeat their answers, which send 80% of them to dialing zero immediately, thereby canceling any automation cost savings and increasing caller frustration. Also, the word “um” breaks most automated speech recognition systems.
  • Confusing menus Callers don’t always pay attention to every menu choice or may not be able to figure out which choice will get them to the right queue.
  • Lack of CTI Computer-Telephony Integration allows the voice channel (IVR) to interact with digital databases, such as the CRM database and the agent screen. If a company hasn’t installed CTI, the caller’s utterances will be used only for IVR routing, and the caller will be forced to repeat his information to the live agent.

We have taken an innovative approach to IVR innovation: we combine machine automation with pinpointed human intervention where needed. The Spoken Smart IVR has a human Silent Guide that monitors up to ten simultaneous calls, just in case there is an unclear utterance. Then, instead of asking the caller to repeat a response, the system continues to the next question, while the Guide quickly makes a real-time correction to the utterance. The result? The caller never has to repeat a response and gets to the agent faster.

Don’t believe it? View a demo of the Spoken Smart IVR and tell us what you think!

Speech recognition still has a ways to go before every caller is automatically understood every time. In the meantime, we believe in innovating the solutions that do exist to provide the best IVR experience possible.

Related Posts Plugin for WordPress, Blogger...

Leave a Reply

Your email address will not be published. Required fields are marked *