top of page

Speech Recognition – Enhancing the Power of IVR – from Security to Sales

2019 is the year that speech recognition enhanced telephone services will step out of the shadows and take up their rightful place as an integral part of traditional IVR services. No longer will speech recognition feature in just a handful of niche applications. Not only is the time and technological climate right but the public expects secure Speech Recognition services to be available via their phones.

Amazon’s Alexa, Google’s Assistant and iPhone’s Siri have made staggering inroads in the public’s psyche. Alexa alone has more than 26,000 voice apps on offer; customers can shop using Google Assistant (Argos is one such example); 50% of all web searches will occur via Voice by 2020 (; and with the over 50s owning the greatest number of smart speaker devices, Greenlight estimates that by 2025 this age demographic will account for as much as 75% of all retail spending. Companies are correct in their thinking, therefore, that speech recognition is a prerequisite to enhance IVR services.

The maturing of existing speech recognition services from the likes of Nuance and LumenVox as well as network improvements in speech quality, mean that even higher levels of accuracy are being achieved. Last year alone, accuracy levels exceeded 95% – that’s the same quality you’d expect from two adults communicating (teenagers excluded, of course!). And increased accuracy means increased security: systems can now use an individual’s unique voice characteristics to prevent fraudulent transactions.

Nuance, confirm that surveys repeatedly show that customers prefer this voice authentication over more traditional pins, passwords and questions.

So why should you incorporate speech into IVR? Well for a start don’t be left behind and out in the cold. Remember how the astonishing impact of mobiles was underestimated such that many companies continued to plough money into website offerings, rather than into app’ development. Obviously, it’s not good to repeat the same sort of mistake. But what can Speech Recognition do for you? Well, it can help realise the full potential of your IVR services with enhanced security, return on investment (ROI), improved customer service levels… We’ve outlined just a few of those benefits:

  • ROI – the big one. Automated interaction is likely to cost at least 25% less than live agent services.

  • Significantly reduces the risks of fraud. Speech can verify a caller’s identity by comparing his or her voice with stored voiceprints. This security has a positive outcome for more vulnerable callers, who may, for a variety of reasons, struggle to pass security with passwords, pins etc.

  • Avoids penalty charges such as non-compliance fines. Speech Recognition Analytics obviates the need to spend valuable man-hours physically listening to recordings and therefore risk being exposed to sensitive data such as card account numbers etc.

  • Adds a further layer of security for ID&V (Identification and verification). Speech can also gather more information such as number plates, postcodes, account numbers with alphas present etc which DTMF (touch-tone) cannot capture.

  • Intelligent routing through predictive analytics quickly routes call types to the most suitable agents. An agent’s skill, as well as interaction capability, can be selected to deal with a particular client demographic. This minimises the number of unnecessary callbacks as well as shortening call times due to improvements in first call resolution rates.

  • Real-time monitoring, along with existing call data (i.e., call outcomes, buying history, level of spend etc) delivers directions to agents to maximise each call’s potential. For instance, customer lifestyle tracking helps provide insights to identify opportunities for post-sales follow-ups, cross-selling etc.

So, what’s wrong with DTMF alone? By its very nature, DTMF presents limited menu opportunities. How often have you listened to an IVR menu only to decide that no options are suitable, except perhaps hanging on for a year and a day to speak to an agent? Ambiguity is undoubtedly a feature of DTMF too. Callers come in all shapes and sizes and by definition couch their requests in all shapes and sizes even though these often boil down to exactly the same thing. Speech, however, can overcome these difficulties through AI ‘natural language’ recognition. With techniques such as ‘word spotting,’ your IVR system would only need to recognise the word ‘accounts’ for instance, in order to direct a caller immediately to your accounts’ department.

From security to sales, it goes without saying that using Speech can really enhance your IVR’s performance capabilities. Capturing and understanding non-numeric input means that agent hold time is reduced and the “IVR Maze” of going round and round in ever-decreasing circles never being quite understood, is eliminated. In short, you get increased customer satisfaction levels, increased security and, by definition, better returns on investment simply by getting to the point faster.

Call us today on 0161 537 7707 or email us at

Post: Blog2_Post
bottom of page