Does Automated Speech Recognition Limit the Growth Potential of IVR Technology?

Self-service has become very pervasive in customer service. We’re getting used to it. We scour Web pages and FAQs for answers, we search user groups, we initiate Web chats with agents and, when we still can’t solve our own questions, we get on the phone and begin the self-service all over again. While interactive voice response (IVR) technology has come a long way in recent years, it’s still not customers’ favorite method of self-service. It’s not necessarily the fault of the technology, however.

“Customer dissatisfaction with IVR systems is largely the result of a mismatch between customer needs and expectations and the ability of a company’s customer care resources to meet those needs efficiently and consistently at high volumes,” Dan Miller (News - Alert), senior analyst at Opus Research, recently told the Web site Business2Community. “What the customer requires often does not map directly to the rigid set of menu options offered by a traditional IVR.”

It’s true that IVRs were once very rigid: about as flexible as bricks. They didn’t fit call center needs very well, nor did they fit customer needs. But does that need to be the case anymore, given the advances in cloud-based, simple-interface tools?

While things may have improved on the call center side, Interactions Corporation’s Dan Fox, writing for Business2Community, says that little has changed on the customer side, and it’s all because of the limitations of automated speech recognition.

“Why has IVR development slowed to a crawl? Because the core of IVR is Automated Speech Recognition (ASR), which is a stunted technology,” writes Fox. “ASR has been promising Natural Language Understanding for years, decades even, yet we have not seen any major developments. Even understanding numbers over the phone with ASR is a challenge. And alphanumeric strings like serial numbers? Forget it.” Fox notes that even Apple’s (News - Alert) Siri, touted as a next-generation speech technology solution, has her moments of confusion.

So how do you build a superior IVR experience with a “stunted technology?” Fox says it’s all about designing automated systems that aren’t shackled to a strict call flow: one that allows callers to engage using complete sentences and natural language.

“A truly resilient speech-based automated system that held conversations with customers in automation the same way they engage with a live agent,” writes Fox. “Powered by breakthrough, advanced artificial intelligence, virtual assistant technology fulfills those lofty goals.”

In other words, call centers interested in reducing customers’ frustration with IVR need to seek out technology that blends automated speech recognition and text-to-speech rendering with a human touch so that interactions with the speech-enabled IVR system feels as natural as talking with a live CSR (News - Alert).

Edited by Rachel Ramsey