In today’s business landscape, data is power. Data enables companies to interrogate their own operations, glean game-changing insights, and when paired with AI, make impactful predictions that can help pave the way to success.

While predictive analysis has long been one of the most used and sought-after use cases for AI in business, its potential is being supercharged as AI and cloud technologies improve. Now as organizations increasingly integrate AI into their customer service processes, it’s unlocking a whole new crop of customer-focused data that can not only transform customer service for the better, but also make predictions that can help propel and shape the entire company.

A 2023 Gartner survey revealed that 79% of corporate strategists believe that AI and analytics will be critical to their success over the next two years. Even going back to 2018, Harvard Business Review called the use of AI for prediction “analytics on steroids” and identified it as the second most common use of AI across businesses it surveyed.

“Cognitive insights provided by machine learning differ from those available from traditional analytics in three ways,” wrote Rajeev Ronanki and Thomas H. Davenport, both thought leaders focused on AI and business analytics, in Harvard Business Review. “They are usually much more data-intensive and detailed, the models typically are trained on some part of the data set, and the models get better — that is, their ability to use new data to make predictions or put things into categories improves over time.”

Here’s a deeper look at the role AI and data — especially customer service data — can play in helping companies predict everything from ROI and churn to specific customer service issues.

“The voice of the customer should be considered in all parts of the business, and insights from AI-powered customer service is a great way to champion that voice,” said Krystal McCann, senior director of professional services at Ada.

Using AI to predict best-selling products, where to put them on the shelf, how to market them, and everything in between

Companies are using AI prediction to inform every step of the product journey from conception and development to sales and marketing, and then later, for analyzing how all of those AI-led decisions actually played out.

Stephan Gans, chief consumer insights and analytics officer at PepsiCo, once told me for an article I reported in VentureBeat how the company has found success using AI to uncover the types of insights consumers don’t report in old-fashioned focus groups. By deploying AI tools to analyze online social conversations about food preferences, they’ve actually been able to create products consumers didn’t even realize they wanted — and that went on to become top-sellers. One example is the company’s Off The Eaten Path seaweed-flavored snacks.

“If you would’ve asked consumers, ‘tell me what your favorite flavors are and let us know what you think would be a great flavor for this brand,’ nobody would have ever come up with seaweed…But because of the kind of listening and the outside-in work that we did, we were able to figure that out through the AI that’s embedded in that tool,” he said.

But the prediction doesn’t stop there. After using AI to inform what products will likely deliver a successful ROI, PepsiCo also started using it to predict exactly where the product will sell best and how to sell it. The company uses AI to predict and target not just the most optimal regions for selling a product, but which specific stores and even which exact shelves. The marketing team also uses AI to create hundreds of different versions of ads for a single product, each predicted to best reach a different audience segment.

This is all just one example of how countless companies are using AI to make predictions to power their businesses, and there are plenty more industry-specific examples. Travel companies like Hopper and Expedia report using AI to predict travel trends and inform business decisions, and executives at retailers like Kohl’s and Walmart have spoken about AI as a gamechanger for making predictions about and optimizing their supply chains, for example.

Just last month, Walmart unveiled a host of new ways the company is using AI to analyze what needs attention in its stores and improve efficiency. Powered by augmented reality, store associates can now use an app to scan an entire wall of boxes to see if they contain any items that are out of stock. Additionally, they now use a predictive algorithm to determine the best way to organize shelves based on customers’ buying patterns.

“It has helped our team move at a very exponential pace, to make sure that we’re protecting our sales floor,” Josh Strudl, store manager at the Secaucus, NJ location where the company recently showed off the new AI capabilities, told Semafor .

Using AI to predict customer service issues before they become problems

The predictive capabilities of AI can work wonders for the customer service side of businesses, too.

"Another trend that we are likely to see in AI-powered customer service is enhanced predictive capabilities,” reads a 2023 Hubspot report on the future of AI in customer service. “As AI becomes more sophisticated, it can predict customer needs and behaviors with greater accuracy, which could improve customer experiences and increase customer satisfaction."

BMW, for example, announced in September that it created an AI system to analyze customers’ vehicle data, predict their service needs, and proactively reach out with a plan to help. Rather than the customer being blindsided by an issue with their car and having to reach out to get it serviced, the prediction capabilities of AI-powered data analysis is allowing BMW to do the heavy-lifting for its customers. This not only makes customers’ lives easier, but offers a great customer experience that’s bound to make them feel like the company truly cares and has delivered value.

Hutch Morzaria, a customer experience director who helps companies deploy AI for customer experience, said in conversation for this article that as data continues to accumulate across businesses, these predictions will definitely become more prevalent in customer service.

He’s helped organizations use AI sentiment analysis to analyze the tone of customer service cases and prioritize certain tickets, allowing the customer to get to an agent quicker. He’s also used AI to triage customer service work and direct issues to the agents with the most relevant skills and knowledge.

Similar efforts to predict and stamp out customer issues are growing across the banking and financial industries, too. Using AI to detect fraud and reach out before the customer calls in a panic, banks are coming across as prepared and ready with solutions — setting up customer interactions for success. Companies across industries are also using AI to shut down attempted cyber security breaches, prevent service outages, and preemptively put a stop to other issues that would cause a wave of frustrated customers (both consumers and businesses) to call in.

Using insights from AI-powered customer service to make predictions that power the whole company

When done well, AI-enabled customer service transformation “can unlock significant value for the business — creating a virtuous circle of better service, higher satisfaction, and increasing customer engagement,” according to a 2023 McKinsey report . By increasing customer engagement, it says AI-powered customer service can result in increased opportunities to cross-sell and upsell customers, while reducing the cost-to-serve.

A recent report from Gartner echoes this sentiment, stating that while many service leaders jump to the cost savings potential of AI, “one of AI’s key benefits is in its ability to obtain insights and predictions.”

Ada’s case study on Indigo , a Canadian book retailer, provides one example of how data gleaned from AI-powered customer service can inform the business at large. Analyzing insights on delivery delays and how often customer orders required customer service attention revealed inefficiencies and enabled the company to streamline its processes for onboarding new delivery carriers, establish minimum standards, and clear data-sharing requirements up-front.

“With the confidence gained from having a complete, centralized view of the company's delivery network and every package moving through it, the Indigo team was able to make operational changes organization-wide,” the company told Ada.

Even the state of New Jersey has been reaping the benefits of data from AI-enabled agents. An IBM case study with the state’s Department of Community Affairs details how the company’s technology helped the agency better connect with citizens in need and has “been critical in terms of business and executive management.”

“We know how many calls are coming in. We know when the busy times are, which helps with staff allocation. We know when our average call times are going up or down and can respond as needed,” said John Harrison, director of information technology for the Department of Community Affairs (DCA) regarding its program to help citizens who have fallen behind on their energy payments. “We even saw a dip in applications at one point, so we knew we needed a new mass email outreach campaign to get the message out there again to the public.”

Exactly which insights from AI-powered customer service can be most useful for business depends on what you’re trying to achieve, and they can help companies in different ways. But one especially valuable insight is proactive notifications of customers at risk of leaving, according to Morzaria, which are critical in combating churn. He said that feeding a customer’s overall sentiment and activities back through AI can elevate patterns and help a business determine whether or not a customer is at risk and might be shopping around for another vendor.

Overall, these capabilities can not only simplify and streamline analysis, but unlock insights that simply weren’t possible with traditional methods.

“What’s particularly noteworthy is the acceleration of this continuous improvement due to availability of AI-driven customer service insights,” McCann said. “Improvements and iterations that may have been on the order of months to identify and form a hypothesis around now may take form in weeks, even days in some cases.”