Creating a Powerful, Data-Driven Customer Service Organization: Part 2

As I wrote in last week's column, efficiency and proactive, personalized customer service are not mutually exclusive. Augmenting customer service capabilities with machine learning can help agents work faster and smarter, enabling support organizations to increase productivity, cut costs, and nurture loyal customers.But the power of machine learning isn't limited to customer service; it holds the potential to deliver great outcomes across the entire organization.

Data about your customers is becoming increasingly interconnected but ever more challenging to sift through for decision-making. While there's growing awareness about the importance of serving customers well after the sales process, customer support is still often viewed as a reactionary, efficiency-driven cost center. By augmenting human decision-making with machine learning, you can be systematically data-driven and proactive in revealing critical customer insight. The payoff? This creates an opportunity for you to bring customer service to the forefront of conversations with other functional areas and deliver value to the rest of the organization.

Machine learning: making customer service more proactive

To set the scene in case you missed last week's article, machine learning teaches computers to think and act like people. It analyzes past interactions to understand what led to great outcomes, both inward-facing, such as who is the best agent to work on this customer issue?, and outward-facing, like what sales and marketing activities (or decisions) lead to customers requiring little to no support? Machine learning then recommends the next-best action to influence a successful outcome for each customer interaction, all while updating its recommendations based on the latest outcomes without the need for manual input. As an intelligent layer on top of existing systems, machine learning helps merge customer service activities and data with CRM, marketing and product data.

How a customer churns, how loyal he is, or whether he turns into a high-value customer all depends on how he came in, how he on-boarded, and how he was treated. It doesn't make sense to approach sales, marketin,g and support as separate elements. Yet often data and organizational silos make it impossible to link together these critically connected phases of the customer lifecycle.

Machine learning helps you go backwards through time to extract patterns from a fuller range of customer data, to tease out and discover all those interactions, like what made a customer want to buy, or why a new product feature or campaign led to a spike in support calls. This is where real engagement and better outcomes happen.

Enabling customer service as the hub of customer success

You're already on the frontlines of retention. Now it's time to be a proactive, data-driven leader. By using insight gleaned from machine learning as a lever to lead conversations with other parts of the business, you can serve as the hub of customer success, driving change that keeps customers engaged and happy.

Let's start with product development. Machine learning could reveal that agents are fielding an overwhelming number of tickets for customers getting tripped up on a certain step in the onboarding process. This is your chance to weigh in and partner with the head of product to make onboarding more seamless. Identifying where confusion exists allows customer service to be proactive in resolving that confusion, while also informing future product direction. Now a partner to the product team, you're supporting the business in driving better retention rates.

The same goes for marketing. Traditionally, marketing and customer service have lived in their own separate worlds, with primarily anecdotal feedback going back and forth. Machine learning turns this qualitative information into usable, quantitative data that can be directly aligned with campaign development and measurement. For example, it might uncover that a new marketing campaign is generating customers with common questions that can be proactively addressed to head off issues; or that customers with specific characteristics have extremely high support needs. As a result, you can be more engaged with the vice president of marketing, helping him improve campaigns in ways that directly align with customer success.

By enabling you to funnel what you know back to the organization responsible for revenue, including the CEO and CFO, machine learning is also starting to play an important role in the evolution of customer service from cost center to innovation center. Continuing the example from above, machine learning can help you bring data about high-support customers to inform the acquisition and marketing strategy, ensuring the cost of support is factored into ROI calculations. At the same time, machine learning can identify opportunities to turn support tickets into revenue opportunities, either by using support agent resources in cross-sell/upsell opportunities (or transferring them quickly to a sales team), or pinpointing where support might be a revenue stopper and prioritizing those issues.

As customer service leaders look for new ways to serve as a proponent for customer initiatives in the rest of the business, the tide is turning in how the enterprise views them and the support organization. Powered by machine learning, your organization can transform from a reactive sounding board to a live feed proactively shaping the entire organization.

Jeff Erhardt is CEO of, which delivers predictive applications built on machine learning technology to simultaneously optimize customer service operations and deliver an exceptional customer experience. Previously, he was chief operating officer at Revolution Analytics. He holds an engineering degree from Cornell and a masters in business administration from the Wharton School.