5 Ways AI is Disrupting Contact Centers

For a painfully long time, business leaders treated contact centers as cost centers. They assumed customer service was inherently unprofitable and squashed costs wherever possible. But the days of the cost center as a contact center are coming to a close.

 

Head to any customer service, customer experience, or marketing publication—and the change hits you in the face: analysts and vendors are now decoupling the “cost center” culture from the customer service center and repositioning it as a profit center.

 

These pieces represent a seismic shift in the industry—an entirely new way of thinking about and deploying contact centers. They argue that contact centers can generate revenue, not just cost money. While the move is encouraging, we’re still in the early days of the transition. It’s all well and good to talk about contact centers becoming revenue-generating monoliths—but what does that mean?

 

To explore what the future holds for profit center contact centers, I’ve identified five new and emerging technologies that will drive the transition from cost to profit center.

 

All companies want a single view of their customers—but creating that view is difficult. Customer interaction happens on many different mediums, including live chat, email, phone, and social media. In the past, companies relied on agents to manually document their disparate conversations on a central platform. But condensing an hour-long conversation into a 50-word summary lost a lot of detail.

 

Now, we have integrated systems that can collect all that data. It can log text data from SMS and emails and automatically transcribe voice data from video calls and voice. New technology means organizations can have a record of everything, not just a 50-word summary from their agents.

 

Most agents can innately tell if a customer is happy or angry, satisfied or frustrated, engaged or distracted. But it’s hard to scale and operationalize those insights. That’s where natural language processing (NLP) comes into play.

 

With your new dataset of all customer interactions, you can deploy NLP to determine the sentiment of customer interactions and extrapolate the customer’s emotional state. The analysis works on three layers:

 

 

There’s a lot you can do with sentiment analysis, but the headline advantage is predictive customer churn . Usually churn is backward-looking. A customer leaves and then they become a churn statistic. But when you understand sentiment trends, you can highlight at-risk customers before they cancel their contracts.

 

Say Acme Corp’s sentiment—both for individuals and the company—began trending down. There’s clearly a problem, but it hasn’t caused Acme to churn—yet. Sentiment analysis grants you time to investigate the root cause: Is there a product fault causing dissatisfaction? Are the end-users struggling to use your product? Are there local or geographic issues at play? Whatever the specifics, sentiment analysis acts as a trigger, sparking proactive work on your side to save customers.

 

Bill Gates once said: “Everyone needs a coach. It doesn't matter whether you're a basketball player, a tennis player, a gymnast, or a bridge player.” He’s totally correct. Coaching helps people improve everything—their outlook, their performance, and their results. But coaching is expensive and it’d be prohibitively expensive to employ personal coaches for every contact center worker. But what about an AI-powered coach? That’s a different story.

 

Because of the tremendous increase in computing power, we can now run transcription in real time. Combined with strong NLP services, we can mine the live transcripts for coaching opportunities and deliver tips, advice, and corrections to agents in the moment.

 

Consider automated compliance.

 

For highly regulated industries like finance, there’s a lot of red tape. Until agents know the industry like the back of their hand, it’s easy for them to skip mandatory steps or miss necessary information. Say you have a six-step identity verification process. With an AI coach, you can monitor what the agent is asking, ticking off each step as they progress. If they miss one, you can trigger a warning on their screen: “Hey Joe, you haven’t completed identity verification. Please restart the process.”

 

It’s like an always-on coach or co-pilot. The AI-coach rides along with agents, helping them improve their performance on every call and through every shift by collecting important metrics. These metrics, such as sentiment analysis, help agents improve their skills over time, helping to cut down human error.

 

Ten cents—that’s how much it costs to field a self-service query online. What about a call to a contact center? It costs around $12 to field a technical support query. That’s a 120X difference.

 

But here’s the problem: Historically, knowledge management systems have not been great. Here’s a real-world example. Go to the IRS website, type in the quote “When will I get my refund?,” and hit search . What comes up? Nothing, right? The crazy thing is that the page (Where’s My Refund?) exists. When folks can’t find answers on their own, they turn to contact centers, tying up human agents with often simple fact-finding requests.

 

Thankfully, search tech is improving. We’re replacing dumb keyword searches with intelligent semantic searches. Instead of looking for your exact search keywords, semantic searches understand your query and pull in the relevant information. This goes far further than just smart knowledge bases, too.

 

Imagine you want an update on your delivery from an online retailer. You ask the website’s chatbot, “When will my order be delivered?” Instead of directing you to a delivery FAQ page, the search grabs the delivery status from your order and responds with an exact update: “It’s out for delivery. Your driver should be with you between 2 pm and 4 pm.”

 

With effective knowledge management, we can free human agents from tedious and simplistic work. Instead of looking up one delivery status after another, they can focus on the trickiest, highest-value queries.

 

Today’s consumers want to interact with companies on their terms via their chosen channel, for their individual purpose, on their schedule. The practice is called omnichannel communication and it’s pretty much table stakes at this point. While there are omnichannel communication platforms that simplify delivery, few support AI-powered conversations. Instead, you manually designed Messenger bots and WhatsApp decision trees.

 

But we’re starting to see the end of this siloed back-end approach. Now, we have platforms that allow you to design conversations centrally and deliver them to whatever channel you want—Messenger, WhatsApp, SMS, live chat, whatever.

 

Not only does this deliver a better customer experience (folks can interact wherever they’re most comfortable), but it also helps companies collect more data. You develop a more rounded picture of each customer and demographics. In turn, that informs marketing, sales, and customer success, helping refine every element of your go-to-market strategy.

 

In his article for Forrester, Principal Analyst Max Ball said it best: "We all owe it to our customers, our organizations, and ourselves to look at contact centers through the right lens to provide the best, most cost-effective service."

 

For too long, we’ve ratcheted up KPI-based customer service expectations while cutting budgets to the bone. It’s time to rethink the contact center and empower customer service to do more than merely answer simple questions. The transformation is already well underway. What matters now is how you respond: Will you stick with the status quo and hope everything blows over. Or will you act early, adopt a profit center mentality, and lead the charge to a better role for contact centers?