Search
  • juanitarollins

Speech Recognition Software

Updated: Feb 8

What every Contact Center needs it to efficiently transcribe calls and identify trends.




In a 2018 study conducted by Microsoft Corporation on the state of customer service globally, they found that 44% of people surveyed in the United States prefer phone or voice to any other customer service method. That’s an important stat for Contact Centers, especially when you also consider 23% of people prefer the live chat option available on many sites. In other words, most people want to interact with a live human. While chat transcripts are easy to pull, call transcripts are more than a little harder. Therefore, every Contact Center needs an effective and efficient way to transcribe agent-customer calls without having to manually listen to and manually transcribe every call manually. Can you imagine the time commitment that would take for even a small Contact Center?



One of the best ways to effectively transcribe calls and capture this data is through Speech Analytics (often simply called “Speech”), which is a process of automatically transcribing recorded calls into text and then analyzing them for key terms and phrases. With speech recognition software, a Contact Center can identify possible trends by looking for the number of times terms are being used. Speech recognition software can analyze a tremendous number of calls in just a few hours. Speech can also be used to identify different types of calls based on the key words spoken (ex. When calling a bank to check your balance the system can look for phrases like “account balance” or “balance inquiry”). With Speech Analytics, any Contact Center now has the power to turn raw customer interaction data into relevant, usable insights through automated processing. The Value of Speech Analytics Speech Analytics can be applied to all calls in a relatively short time. Meaning, you don’t have to wait weeks or months for the data, when it can be curated within just hours. Not only is data available faster, but you can also have data on all calls rather than a select few. From this data, trending analysis can begin. This is where the true value of Speech starts to show itself.



As Speech systems work, they search for established and unique trends in key words or phrases. This can help identify potential issues and even root cause analysis of those issues. It is highly recommended to have at least 40,000 transcribed calls before the trending analysis is considered valid. When Speech systems identify trends, they start looking at the other words and phrases used near those trending key terms/phrases to help look for possible root causes of the trends themselves. This is helpful not just to know what is currently trending, but also provide possible insight into why they’re trending. Speech Analytics and Quality Management Though Speech Analytics is a powerful tool, and one every Contact Center can benefit from, it does have some limitations.



First off, the technology is relatively new so it’s not without error. Second, no matter how precise you are in defining what phrases you want it to actively look for, some phrases will get missed. This is because a customer may be using different words or phrases to convey the same idea or problem than you and your team may have thought to look for. Yet, even with these limitations, Speech Analytics is an invaluable tool for Contact Centers to use to enhance their QM processes.

_______________________________________________________________ How would having all of your calls automatically transcribed for you affect your QM processes? Could your Contact Center benefit from almost real-time insight into why customers are calling? Would you like a faster and more thorough method of automatically identifying what types of calls have come in? CLICK HERE to schedule a short call with us to discuss how we can help bring Speech Analytics into your contact center.


12 views