How to Leverage Predictive Analytics in the Contact Center

Learn more about predictive analytics at SpeechTEK 2022.

See more videos from CX Connect 2021 on the CRM YouTube channel.

Read the complete transcript of this clip:

Terri Kocon: We all know that channels of communication and channels of data that we get from customers are just going to continue to increase. We have devices with internet access built into them. That's essentially everywhere. Alexa is everywhere. And these are all key channels of data. So it becomes even more important that we have tools on top of this data that will allow us to truly understand what's happening within the contact center.

So I'm going to break this down a little bit, and I'll give you some examples of how artificial intelligence and machine learning are used to provide predictive and prescriptive analytics that can really help you help a contact center start to take proactive instead of reactive measures and be able to perform things like targeted quality scoring instead of random analysis and help to spot agent performance and customer satisfaction trends without having to to look for them. They'll be spotted automatically.

So, to give you an example of machine learning technology, I'm going to speak from the context of the product that I'm familiar with, which is obviously the Calabrio product. And one of the key technologies that we use machine learning for is our predictive capabilities. So, specifically around predictive evaluations and predictive NPS (Net Promoter Score). The way this works, we use our source data, which would be recorded interactions. And we're going to combine that with all of the metadata that we have around those interactions. And that includes metadata from the ACD, from the CRM, any other sources that we might be incorporating into that contact. And then we layer that with the analytics data that we have, including speech analytics, text analytics, desktop analytics, things like talkover events, silence during the interaction, et cetera. And so we're taking all of this data that we have in regards to context, and we're correlating that with an evaluation score.

And so now we know what are all of the things that, for example, high-scoring contacts have in common and what are all the things that low-scoring contacts have in common, and everything in between. And then we're able to use that. We're able to extrapolate out for 100% of our contacts, what a predicted evaluation score would be. Or, for example, with surveys for predictive net promoter scores, if it's survey data with an NPS question, we do the same thing. And so we're able to take a small sample of manual survey contacts and use that to correlate what high-scoring contacts and low-scoring contacts have in common, and then use that to extrapolate out, to 100% of our contacts, what a predicted promoter score would be. So again, if a contact center is lucky, they can manually review about 2% of their interactions.

And then we're able to take that 2%, and by applying machine learning concepts, we're able to extrapolate that 2% of evaluations or 2% of survey data across 100% of interactions. And then the contact center will have a more holistic, more complete picture of how it's performing overall. So that's one example of machine learning technology and how it works.

Another example of AI in the contact center that also leverages machine learning is sentiment analysis. Here we are trying to recreate the thought processes that a human might go through, and that's using an algorithm for concepts like emotion or sentiment detection. That's a much more complex question. How is the caller feeling about this interaction? What is the sentiment? Is this a positive interaction, a negative interaction, or is it a neutral interaction?

Is there anger being displayed? What are the different cues that the AI engine could pick up on in order to give us an idea of the emotional content or the sentiment of what that particular contact might be? And again, this is an extremely powerful tool that allows us to really understand one of the most important questions that we have to ask as customer service professionals: Are we delighting our customers? That should be a question we're asking ourselves every day. And emotion detection and sentiment analysis helps us to get there.

Related Articles

The Benefits of Automated IVR Training

Bespoken Chief Evangelist Emerson Sklar outlines the key advantages of automated IVR training, such as increased call center ROI, in this clip from his presentation at CX Connect 2021.

CRM and the Three Pillars of Data Management

Validity's Jeff Foley discusses what sets companies with higher-quality CRM data apart from their competitors in this clip from his presentation at CX Connect 2021.

Goals and Challenges of IVR Modernization

Bespoken Chief Evangelist Emerson Sklar outlines the essential goals of modern, AI-powered IVR--reduced cost, increased throughput, and improved customer satisfaction--and the challenges organizations face in reaching those goals in this clip from his presentation at CX Connect 2021.

How Empathy Improves CX and Drives Outcome-Based Selling

Concentrix' Andy Bird and Will McCain discuss how an empathy-based approach to customer service can drive differentiated experiences in this clip from their presentation at CX Connect 2021.

How to Deliver Frictionless CX

Validity Senior vice President of Marketing Kate Adams explains how organizations can remove friction from the customer experience through self-assessment and self-secret-shopping and analyzing their CX flow from an external perspective in this clip from her presentation at CX Connect 2021.

Key Customer Service Challenges Solved by IVAs

SmartAction's Brian Morin and TechStyle Fashion Group's Aarde Cosseboom discuss service challenges many customer-facing organizations face, and how TechStyle Fashion Group mastered them with intelligent virtual assistants in this clip from their presentation at CX Connect 2021.

Key KPIs for Outbound Call Centers

First Orion's Sara Hurst and Kent Nicholas explain how to deploy the right KPIs to measure the success of outbound call centers and branded customer communications in this clip from their presentation at CX Connect 2021.

The Case for Conversational AI

SmartAction's Brian Morin and TechStyle Fashion Group's Aarde Cosseboom discuss how TechStyle Fashion Group made CX more manageable through conversational AI and sold stakeholders on the solution in this clip from their presentation at CX Connect 2021.

The Challenges of Managing Customer Interactions

Verint Customer Engagement Solutions' Daniel Ziv offers best practices for meeting the myriad challenges of customer interaction management--particularly as volume demands increase and workplaces become less centralized--in this clip from his presentation at CX Connect 2021.

Who Benefits from Contact Center Analytics?

Calabrio Product Marketing Manager Terri Kocon explains how data gathered through contact center analytics can benefit organizations in areas well beyond the contact center itself in this clip from her presentation at CX Connect 2021.