Dream Bigger with Real-Time Speech Analytics

If you have watched Sylvester Stallone’s blockbuster sci-fi film, Demolition Man (1993), you’ll remember the classic scene:

A machine on the wall fines Sylvester Stallone’s character every time he says a bad word, announcing, “You are fined 1 credit for a violation of the Verbal Morality Statute” and printing a ticket of fine. If you are thinking of a relationship between the title and this intro, spoiler alert, this scene represents a basic application of real-time speech analytics in real life.


Image credit: https://media.giphy.com/media/uRpmGfRwj7ZXa/source.gif

This example may be found in 1993, but real-life industries have adopted these types of technologies for the last 10 years, and speech analytics applications are no exception. Because customer experience is everything to keep a customer happy and work with you, companies are keeping their eyes on every technological improvement in the industry not to be left out from the race. Therefore, you need to understand your customer, and that road leads you to analyze your customers. By 2021, more than 40% of all data and analytics projects will relate to an aspect of customer experience. By 2022, more than half of all major new business systems will incorporate continuous intelligence using real-time context data to improve decisions. So, you need to gather data and act quickly to give an enhanced customer experience and lead the way.

Using speech data to improve business decisions is not new. A lot of companies use speech analytics to dig out valuable insights within their contact center. Post-call speech analytics applications are good; they are reliable, fast companions. But what about living in the moment? What about taking action on the call? What if it is too late after a customer hangs up the phone with a bad experience? So, these questions lead us to a way that real-time actions are too important to overcome different situations on-the-go.

Post-call speech analytics applications analyze the voice after a call is hung up. Real-time speech analytics (RTSA) applications analyze the voice while the conversation is going on between your customers and agents. That means now you know everything in every conversation in your contact center. From now on, you can dream big and take actions just like they did in “Demolition Man.” You can create an online coaching platform that assists your agents in real-time when they are on the call or create alerts based on specific customer attitudes.

On the other hand, there are essential to-do’s of RTSA to follow carefully to deliver a successful project:

Define your target: Creating a pop-up every time a customer says something is nice but consider if it is worth to invest on it and if it will save you money. Even sometimes making a real-time action may not be the best answer. So, create a scenario that real-time intervention or action is essential, makes your work easier or saves money such as re-shaping agent behaviors in real-time within churn situations. The scenarios that are considered high level and superficial often result in poor end value for RTSA, so consider every aspect of this project in detail.

Clarify the actions: Within these scenarios, you will need to clarify every action depending on certain situations. Whether you only want to pop-up some recommendations to your agents’ screen or integrate with the customer database to pull out certain offers for your customer, you need to put an action in front of every possible scenario that you want RTSA to take part.

Get a speech analytics application: It is tough to make a real-time speech analytics (RTSA) project if you do not have a speech analytics application. Working with speech data is a new experience for your company. So, first, implement a speech analytics application, get used to working with post- call speech analysis data, and at the end, try to uncover potential scenarios for RTSA investment.

RTSA is an investment, and you need to know why you are spending money on it. Therefore, working with speech data helps you to uncover what your agents’ strong or weak sides are, or what your customers think, do, and say in a conversation.

For example, how often do your agents offer your churning customers an upgrade on their accounts? Is it right to ask this to every customer? What should you offer based on customer mood? How do your customers behave in different moods? These questions and using speech data will lead you to create solid and viable scenarios with real-time speech analytics. Without data to prove, it is very difficult to verify if an RTSA project will be useful or not.

With reliable data to prove that you can prevent churn by implementing RTSA, it becomes easy to convince every stakeholder in your company to engage in this project.

Create a team: One of the biggest challenges that the customers face after deciding an RTSA projectis that there is a big responsibility on the customers’ side. RTSA is a tool that gives you real-time data, and as long as it gives you real-time feed, you need to take actions. RTSA can create pop-ups, trigger certain alerts, or can give you real-time dashboards, but it cannot make a change on your agents’ screens or automate a process on your systems. So that will be your work to take certain actions based on the live data fed to your systems. Sometimes these responsibilities even prevent customers from starting an RTSA project, and it is crucial to know what it takes to accomplish these types of projects. That is why a clear target and carefully designed scenarios are also very important to define the responsibilities in an RTSA project.

So, based on these three steps, we can give you an example like below:

A customer wants to leave your company, and you want your agents to prevent this churn by asking their reasons and offering advantageous packages based on these reasons. First, you need to make every agent ask about the churn reasons to your customers. You must not ask too many questions just like you must not skip asking questions so, track your agents with speech analytics to make sure the churn process is going by the book. You can remind your agents by encouraging them to ask questions or remind them not to ask too many questions by alerting them.

Based on the example above, integrating speech data with a real customer database, finding the best offer, or taking actions on agent’s screens will be important tasks that generally should be carried on by your company’s IT teams. So, adopt a holistic approach while you are designing your scenarios.

After that, capture real-time customer answers with real-time speech analytics, integrate them with real customer data, and find the best offer that fits your customer’s expectations to pop up this offer on your agent’s screen as a link or a popped-up page.

After you follow your steps and dress every agent in a futuristic Sylvester Stallone style, you are all set!

Author: Fahrettin Yılmaz, Business  Analysis & Consultancy Team Leader, Sestek