Blog Home

21 Marketers, Analytics Pros & Business Leaders Reveal the Most Creative Uses of Predictive Analytics in the Call Center


The Team at CallMiner

June 04, 2020

Business man looking at digital screens
Business man looking at digital screens

Call centers are increasingly making use of innovative technologies like artificial intelligence, machine learning, speech analytics, and other data analytics tools to provide more comprehensive support, improve the customer experience, and boost agent productivity and efficiency. Predictive analytics is one such technology leveraged by a growing number of call centers to support data-driven decision-making on everything from staffing to call routing, and even the solutions or offers proposed and the words used to communicate them to callers.

Download our white paper, How AI Improves the Customer Experience, to learn more about innovative uses for artificial intelligence and how you can leverage AI to improve the customer experience.

Companies have just begun to scratch the surface of the potential use cases for technologies like predictive analytics in the call center. To learn about some of the most creative ways call centers are leveraging predictive analytics today, we reached out to a panel of marketers, analytics professionals, and business leaders and asked them to answer this question:

“What’s the most creative use of predictive analytics in the call center that you’ve seen?”

Meet Our Panel of Marketers, Analytics Pros & Business Leaders

Paul Bonea


Paul Bonea is the Founder of Perfect Data.

“An interesting use case of call analytics that we’ve seen is…”

A client in the WordPress space who had a SaaS plugin. When reviewing their support department, they discovered that clients who got in touch by phone had a better overall retention rate than clients who used only emails, who in turn stayed longer than customers who didn’t contact support at all.

In other words, responsive and satisfying support earns customer trust and loyalty. As a direct result of this observation, they simplified the process of transitioning from email to live call. This directly helped by improving the retention rate and lifetime value by around 5-6%. Indirectly, it also got them a lot of good reviews.

Morgan Taylor


Morgan Taylor is a Finance Expert and CMO at LetMeBank.

“One of the best ways I’ve seen to track whether people are…”

Buying off the back of a marketing campaign – and not just determining success based on receiving more calls or not – is by using predictive analytics.

Predictive analytics can track the success of marketing campaigns by using speech mining software to track phrases in relation to it, such as, ‘I saw this offer on your website,’ or ‘I’m calling about your latest TV ad offer.’

Reuben Yonatan


Reuben Yonatan is the Founder and CEO of GetVoIP.

“Predictive analytics provides a great boost for call centers looking to tighten up how they do business…”

One of the most interesting uses I’ve encountered has been using it to predict the success of a follow up call. Predictive analytics has proven very effective at determining customer intent.

By analyzing their buying history, the words and phrases they use, and the number of times they’ve been contacted (or contacted the center themselves), it’s possible to determine if a follow up call will lead to a sale. This helps agents focus on the leads that are most likely to result in a sale.

Nicole Ward


Nicole Ward is the AVP of Corporate External Relations at Synchrony.

“Synchrony is integrating predictive analytics into the call center in a number of ways to improve the experience when customers call…”

One is through a speech capabilities solution we have been piloting since late last year. Using predictive analytics, the tool analyzes the tone, pacing, and hold times of both the agents and customers. It also takes into consideration if either party is talking over the other, how often, and what’s being said. All of this gives service agents consistent feedback on each call.

Additionally, Synchrony’s call center agents use our chatbot, Sydney, to answer questions and solve customer problems seamlessly and efficiently. Built on a machine learning and artificial intelligence platform, Sydney currently handles more than half a million inquires a month.

Andy Peart


Andy Peart is the Chief Marketing & Strategy Officer at Artificial Solutions.

“With conversational AI becoming a key player in call centers, the use of predictive analytics can transform customer experience as we know it…”

The ability to not just inform, but also to notify and help make personalized decisions is a major feature in conversational bots that can be applied to call centers.

For example, machine learning and predictive analysis can assist in evaluating behavioral patterns that can help call centers offer better up-sales, improvements, or problem resolutions by analyzing previous customer behavior and predicting future insights. A call center related to recreational activities can use predictive analytics to forecast weather conditions or peak times to envisage the best moment or day to carry out certain activities. In addition to collected data from past purchases, the final outcome can be fully personalized and tailored suggestions that boost customer experience and customer retention 24/7.

Additionally, predictive analytics in combination with sentiment analysis from conversational AI can enhance customer experience to determine the best approach to specific queries and anticipate significant directions in the customer life cycle. Inbound calls can provide information on how to approach customers according to their dissatisfaction and risk of churning and use predictive analytics to suggest short-term or long-term promotions. Outbound calls can receive insights on the best approaches on pricing and product strengths depending on past data.

Elmer Taboada


Elmer Taboada is the Marketing Manager at DaVinci Tech.

“Artificial intelligence has helped shape the way we handle things…”

With predictive models that use algorithms and past historical data, businesses have increased sales and productivity by predicting future outcomes.

Using predictive analytics for customer retention to renew contracts with existing customers is its most creative usage in call centers. Since expert analysts say that it is 10 times more expensive to find new customers than to keep current customers, using predictive analytics to improve customer retention can increase your profit margin.

A statistic released by McKinsey back in 2014 supports the previous claim and says that businesses who base their decisions on customer data analytics see a 126 percent profit increase compared to those that don’t.

Alex Azoury


Alex Azoury is the Founder & CEO at Home Grounds.

“During the holiday season, we see a huge increase in consumer spending…”

At the same time, a lot of inquiries are just surface level interest.

In one call center, I saw predictive analytics (PA) match the best closers with the best leads. This call center used PA to sort inquiries not only by media channel, but also the topic of the piece of content which tipped them into the sales funnel. In this manner, PA was employed as a lead scorer, helping entry-level seasonal callers automatically prioritize leads. The

benefits were two-fold: Top performing callers were rewarded with – and closed more frequently – the best leads, while newer and lower performing callers were given less qualified leads to practice on.

Rameez Ghayas Usmani


Rameez is a Digital Marketing Expert at PureVPN.

“One of the most innovative uses of predictive analysis that I have seen call centers use is a…”

Metric described as, ‘best month of the year with a high rate of calls picked-up.’ They use the data to find out which month of the year people are most likely to respond to the calls, and then they maximize the use of their resources in these months to make the most of the season.

Jesse Rio Russell


Jesse Rio Russell is the President of Big Picture Research and Consulting.

“Predictive analytics in call centers are most often based on…”

Better understanding caller needs, better directing callers to the right call channel, and decreasing the time it takes to get a caller to the right help. Predictive analytics can be helpful for all of these things, using information already known about the client, gathering more information during call screening, and collecting more information through the call agent.

The best use of predictive analytics I have seen in a call center, however, was not focused on callers; it was focused on call center agents. Most companies are focused on predicting some aspect of caller attributes so that they can increase caller satisfaction with the call. But, turning the predictive analytics machine around, a company can better predict which units and which agents are most likely to produce the best outcomes. Predictive analytics that takes a call center agent attribute as the outcome (such as percentage of calls with positive feedback), the company can better train and support its staff by bolstering the attributes it learned from the predictive analytics, avoid the behaviors that lead to worse outcomes, and can hire and retain the call center agents with the assets that predict better call outcomes.

Alexandra Zelenko


Alexandra Zelenko is a Senior Marketing and Technical Writer at DDI Development.

“Identify who is likely to pay for your products or services with voice predictive analysis…”

You can assess a speaker’s emotional behavior and tone by using advanced machine learning algorithms based on the voice data that was recorded and processed. This math-based assessment of a speaker’s tone is much more reliable than the feeling of most call center agents, as it is based on repeated observations of call outcomes. By measuring millions of calls in a jiffy, the machine learning algorithm can assess your target customers in a few hours. Having that in mind, you can use voice predictive analytics to identify who is likely to pay.

William Taylor


William Taylor is a Career Development Manager at VelvetJobs.

“Call centers are becoming increasingly reliant on predictive analytics to shape their business goals…”

Collecting data is important, but you can only make use of this data if you have robust analytics tools that can identify the patterns in the data. Omnichannel call centers have made the move to advanced data analytics more viable and useful. With an omnichannel

approach, you have a complete picture of all customer interactions and data so you can be confident that your analytics are accurate.

Dashboards that feature key statistics about the call center are becoming increasingly common in the call center landscape. Real-time analytics is also a key part of this trend. For example, agents are using analytics tools to be able to see in real-time what the average handling time is and whether they are meeting their targets. If an agent sees what targets they are meeting and which need some improvement, they can adapt their approach in real-time to get better results.

Bottom line: Predictive analytics help call center agents see in real-time what the average handling time is (as well as other important performance metrics) and whether they are meeting their targets.

Brian Atkiss


Brian Atkiss is the Director of Advanced Analytics at Anexinet.

“I have seen quite a few creative examples of predictive analytics in the call center…”

The most common use case that seems to be very effective is building predictive analytics for customer churn, leveraging speech analytics, text analytics, and combining those results with other structured data fields to identify historical customers who have canceled. This data is then used to build a training set that can then be used to predict future customers who are likely to cancel. While this is probably the most useful example I can think of, another more creative example is from an insurance company that utilized speech analytics patterns to identify certain background noises as potential life events (e.g., a baby crying) and creating targeted offers (e.g., life insurance) for those customers.

Ludovic Rembert


Ludovic Rembert is the Founder and Security Analyst of Privacy Canada.

“Predictive analytics is used to make predictions for a business in the future based on…”

Results that have previously happened and by analyzing data that has been recorded. By blending different techniques to make these predictions, leaders can create valuable insight that can help drive change in areas of the business that need it most in order to retain more customers.

A creative use of predictive analytics is utilizing online security tools to maintain a trusted relationship with customers by honoring their personal privacy and sharing that with them. In call centers, it is hard to make assumptions based on conversations over the phone which can make it difficult to get the customer to come back. But, by highlighting the privacy and security aspects of the company, employees can hook customers with an exciting offer.

Veronika Gladchuk


Veronika is a Marketing Content Writer at SupportYourApp.

“Personalized customer journeys rely on segmentation and prediction models…”

Making it possible to fine-tune the experiences for key customer groups according to their needs. In customer support, this could mean using a different tone of voice, different wording, or making recommendations based on the customer group. This tactic often results in higher loyalty and better brand relationships.

Dan Bailey


Dan Bailey is the President of WikiLawn Lawn Care.

“We of course use predictive analytics to help close sales with business owners…”

Our goal isn’t to be aggressive right out of the gate. We simply give an overview of our services and how it can help them, and if they seem at all receptive, we schedule follow-up calls. Predictive analysis allows us to judge the interest of the client and what we should focus on.

We try to ask about their struggles in finding customers, for example, and predictive analysis helps us create a database of these problems that we can then find creative ways to solve. Many of our clients say their potential customers don’t really see the value in their service over doing it themselves, so that’s something we can work into a strategy on the next call.

In general, it allows us to really tailor the experience for every client while keeping consistent with what we’re trying to do.

Peter Song


Peter Song is a Machine Learning Engineer and Blogger at Haki Review Mashup Company.

“One of the most creative uses I’ve seen is predicting customer’s intention for calling…”

When a customer calls, the person will be guided to enter some personal details. Then, in the back-end, a system can pull the pre-processed personal data such as the goal and the sentiment of the previous calls and chats, activities on the company’s website, the time of the incoming call, and other information. Before the call gets routed to an agent, the system predicts the intent of the current call and displays relevant information on the screen. Armed with this information, the agent can respond more effectively, which also leads to higher customer satisfaction.

Since each call can be ended faster, an agent can handle more cases, thus it can save cost. This case from the customer and all the other incoming calls will be stored and automatically categorized by topic by the machine learning model. As we have more cases recorded, the data can be used to predict a similar intent more accurately in the future.

Conner Cole

Conner Cole is the CEO at Cyber Receptionists.

“The Cyber Receptionists Team is using call center software to predict what our clients’ customers are calling in about…”

By analyzing KPI data like time of day, call tagging, and call duration, our software helps our employees do a better job with customer service.

Creating different types of call tags for our clients allows us to build a data driven analysis of our clients’ calls over time.

Joshua Jones


Joshua Jones is the CEO at StrategyWise.

“StrategyWise helped one $5BN+ company turn a customer support call center into an additional source of revenue by…”

Offering customers products and services at the end of each support call. Using third-party data sources, they built an algorithm to identify customer purchase propensities and recommend unique products for each caller. They then trained the existing in-house software to capture whether the product offering was successful and fed that data into a model that retrained nightly to improve the model’s performance. In order to test and fine-tune the system’s overall effectiveness, a small subset of random calls would receive no guidance, and the sales results would be tested against the model’s results.

In another case, StrategyWise worked with a national energy company to predict call center volume by first predicting root causes of calls. In this case, leaks in gas pipelines were causing customers to call and report the issue. By analyzing factors like the age of the pipe, location, seasonality, weather, and so forth, they were able to dramatically improve on existing estimates of leaks before they occur, which in turn helped the call center better staff and minimize wait times.

Layton Cox


Layton Cox is a Marketing, Sales, and Service Consultant.

“As part of a consulting project for a large IT company, we developed a predictive analytics algorithm that…”

Allowed the collections team to prioritize calls based on the expected success of the collections. The algorithm was based on the client’s historical collections success rate, the economic strength of the client’s industry, and the debtor’s payment history.

Collections identified as a low probability of success were given more focused efforts earlier than traditional collections processes would have indicated. For example, if an invoice was deemed to have only a 25% success rate, the new predictive analytics process would notify the collection team to reach out to the payer for the invoice two weeks before the due date.

Historically, collections only reached out to payers after a due date had been missed. This allowed our IT client to create stronger relationships with their clients by creating payment flexibility for unforeseen circumstances and increase their overall collections rate by more than 20%.

Laura Fuentes

Laura Fuentes is the Operator of Infinity Dish.

“At Infinity Dish, we’ve spent a lot of time curating analytics to further understand the actions and habits of different psychographics…”

What we didn’t realize was that this could actually help us predict how our customer base would respond to the pandemic. By understanding jobs, lifestyle, family situation, interests, etc., we were able to project who would be furloughed, working from home, or still out on the frontline. These insights heavily impacted where we focused our calls. For example, families at home are more likely to invest in satellite TV. Work-from-home professionals would appreciate high-speed internet, while those on the front line would likely resent constant bombardment from companies.

Erico Franco


Erico Franco is an Electrical Engineer and Inbound Marketing Manager at Agência de Marketing Digital.

“We have been using predictive analytics to score leads who contact our sales department…”

Based on their score, our salespeople always know which prospects they should pay the most attention to. We set up the scoring system based on a detailed analysis of the historical behavior of our prospects by CRM and the probability of them closing a deal with our company based not only on contact data such as company size, company sector, etc., but also based on the contact’s behavior on touch points, such as

  • contacts who agree to fill out a second contact form telling more of their goals and challenges, or
  • contacts who are engaged in receiving a call from our team after knowing the price of the service.

Predictive lead scores have proven to be very effective in determining which prospects our sales department should attend to, improving customer experience and sales at the same time.

In what creative and innovative ways does your call center leverage predictive analytics?

Artificial Intelligence Contact Center Operations APAC EMEA North America