Written by Simon Marchand, Chief Fraud Prevention Officer, Nuance

Imagine you’ve decided it’s time upgrade to the latest iPhone. You call your mobile provider, and you instantly feel like you’ve lucked out: The agent you’re speaking to seems great at their job. They find you the perfect plan, and in addition to your new smartphone, they offer to throw in a tablet—completely free of charge.

You don’t actually need a new tablet. But as the agent says, “Hey, it’s free! Do you have kids? You can always give it to the kids to play with.” You agree to the deal, thank the agent, and hang up feeling like you’ve had a world-class customer experience. You even take two minutes out of your day to complete the feedback survey that immediately pings onto your soon-to-be-recycled handset.

Flash-forward a few months. You’re staring at your mobile bill. You’re trying to figure out why it’s consistently $10 higher than you expected. Then you realize what’s happened. Your new tablet—which you’ve literally turned on once since it arrived—wasn’t as free as the agent led you to believe. You had to subscribe to a $10 data plan to get this extra “gift”, something that the agent didn’t feel like it was important to mention.

It’s an awful feeling for a formerly enthusiastic customer. And that should make it a serious concern for any provider that’s competing on experience. (Which, let’s face it, is every provider).

Happily, however, it’s an issue that providers can mitigate with AI-based technology, even as they set about strengthening their broader fraud prevention capabilities.

Driving compliance with agent AI

I’ve written previously about the importance of using biometrics to authenticate customers, especially with more agents working from home. But as well as listening to the characteristics of the voices on any call—to identify the customer, agent, and even known fraudsters—AI can listen to the words that are being said, and guide agents’ actions in real time.

Such “agent AI” can support new agents as they find their feet. It can surface information and materials the agent might need to meet a caller’s needs. And it can provide a timely check for agents who might otherwise be tempted into malpractice.  

Let’s take the imagined interaction we started with. You’ve called your provider to get the new iPhone, and the agent says, “Oh, hey, we’ve a promotion on right now, and I can throw in a free tablet. Is that something you’d be interested in?” This time, however, the provider has an AI-based agent coaching solution in place.

It hears the words “free tablet”, and understands that the agent is offering you a popular promotion. It hears that you’re interested in the offer. Next, it’s expecting to hear the agent explain the terms, conditions, and fees associated with that free device.

When it doesn’t, it sends an alert to the agent, reminding them of their legal responsibility to follow the script, and the topics they need to cover with you before you commit. The agent explains the finer points of the promotion, and the AI listens to make sure you give your clear and explicit consent.

Maybe, now that you have full knowledge of the offer, you decide not to take the tablet. Or maybe you decide to go for it after all. Either way, you make an informed decision, you’re not surprised by your bill, and you’re left feeling positive about your chosen provider.

After the call: targeting agent training

The opportunity here isn’t just to get AI to listen to calls as they’re happening; it’s to get AI involved in auditing conversations after the fact.

Rather than taking a random sample of calls and hoping they’re representative of overall agent performance, AI can help providers to mine call recordings for specific intents and issues. The relevant recordings can then be used to guide targeted training, either for individual agents or for the wider customer care team.

The moment to address challenges—old and new

As providers rush to respond to the fresh wave of fraud unleashed by COVID-19, tackling age-old issues of agent compliance and malpractice may seem like a relatively minor concern.

But if—like so many of the business and fraud leaders I’ve been speaking to recently—you’re seizing this moment to modernize and consolidate your technology stack, it’s important to make the most of new investments. And that means understanding how AI can help your brand prevent criminal activity of every kind.