The Power of Speech Recognition in IVR
Recognition is so much better now than it used to be. We can now just ask our smartphone what the weather is like outside to get an accurate forecast without searching through apps and menus.
With products like Siri and Google Voice search, customers are getting more used to speaking to applications, so it’s becoming increasingly expected from your caller base. With that in mind, wouldn't it be great if you could treat your customers to such a straight forward phone response, rather than making them wade through multiple layers of IVR menus?
Speech recognition systems are now so advanced, they can now engage in "conversations" with you through a technology called natural language programming (NLP). NLP works through the process of machine learning and statistical interpretation, in which software searches through a set of programmed real-life examples to recognise and respond to speech.
NLP works best in fairly simple "conversations" that relies mostly on yes or no answers, or those that have few possible answers. Instead of searching its entire vocabulary for each word in a question and processing them separately, NLP systems react to certain "tagged" words and phrases to respond appropriately to things like "weather forecast" or "pay my bill."
We utilise this natural language tool for quite a few of our clients. So for
instance, if a caller is phoning for a balance enquiry, they hear: "How
can I help you?" They simply say: "I want to hear my balance"
then they are asked for their identification and the system reads their
balance. They don't have to use a main menu, or navigate a couple of other
menus to get down to it.
Voice recognition software gets better by "learning" from each experience. At Eckoh we have a group of experts who know how to interact with speech, we know how to formulate the speech, do the technical side of it so your callers can get in and get out easier than they ever could before. We also have analysts that come through and analyse the calls from real callers are reacting to your application. We call this 'tuning' - where we capture the response and ensure that we understand them. Programmers can add words, pronunciations and grammatical structures that the system is failing to understand. Software is hand-coded with a variety of real-world examples for the software to search and draw from. Average accuracy is around 96% for our clients.
As technology continues to progress, the future of speech recognition software looks to focus on making translation services even more accurate and further developing computers' ability to understand the words they're taking in. We're constantly reviewing the data to see if there are ways we can make it better. We read the industry trends that are cutting edge and move forward with our speech technology to ensure the best performance we can possibly deliver.
For more information click here.
Does it feel like it's panto season in your contact centre every day? If so,…
We've all watched on in horror ... as an out-of-control child creates havoc at a…