Using AI to Drive Human Empathy at Machine Scale

For a few days in early February, the kitten filter Zoom video took the US by storm. Viewed more than 10 million times on YouTube, the short clip was hailed as “an instant classic” by the New York Times for its depiction of a lawyer trapped with a kitten filter during an online court hearing. We laughed because we had all been there with technology–a moment of charming recognition of the challenges and amusements of navigating COVID’s new normals.

In other words, it was a moment of cultural empathy, walking in someone else’s shoes to recognize the familiar challenges that users, shoppers and employees face as they embrace digital products and services every day. But the benefits of empathy go much deeper than a quick laugh and building common ground. For organizations and executives whose growth depends on delighting customers, Deloitte Insights research indicates that empathy leads to stronger financial performance and better customer satisfaction. Turned inward on employee engagement, research shows that 82% of employees would consider leaving their job for a more empathetic organization and 72% of CEOs report that the state of empathy needs to evolve.

But while it is easy to relate to a single funny video shared on Twitter, enterprise empathy requires walking in the shoes of tens of thousands of employees or even tens of millions of customers in order to surface the small triumphs and frustrations that make or break the digital experience for customers. Events must not only be surfaced across billions of interactions, but they must be indexed, classified (for example, does repeated clicking indicate rage or a desirable design element?) and shared by the product, engineering and design users responsible for driving solutions.  The immensity of this data, as well as the need to leverage it across departments and functions, means that AI is required to drive “enterprise empathy,” or human empathy at machine scale.

AI for Enterprise Empathy

Cutting-edge companies are layering AI techniques such as clustering, feature extraction and natural language processing on top of more traditional algorithmic approaches to gain intelligent, prescriptive and empathetic insights that can be rapidly understood and used by business and technical users. Through AI, these companies can now understand and improve entirely new customer paths that were not anticipated or planned for, applying sophisticated clustering, signal/feature extractions, and new heuristics like Rage Clicks (™) to  empathize with and solve customer needs in previously unimaginable ways, as well as model or reconstruct individual sessions on the fly to pinpoint specific errors.

One approach, currently leveraged by diverse companies ranging from Saks Fifth Avenue to Peloton to Zillow, is to use big data infrastructure to support access to massive datasets and AI-driven “augmented intelligence” capabilities that assist with automatic suggestions, contextual comparisons, and visualizations. Once issues are diagnosed, saved report settings and automated alerts allow product and engineering teams to track progress, monitor future incidents and continually learn and optimize together. Enriched human event data can be transformed into data science research and analytics (e.g. via API call or webhook) to create rich insights that inject customer empathy into new processes and operations.

Driving Real Business Value

Enterprise empathy allows organizations to understand and improve digital experience based on real events and evidence, removing blind spots, reducing bias, and driving business value. For example, a company can now identify and prioritize problems they never before knew existed by discovering segments of customers through patterns or anomalies that lead to conversion failure or success. Once surfaced, teams can rapidly act on the most high-value opportunities based on comparative analysis of all segment dimensions and metrics, including the projected number of customers affected and predicted revenue impact. They can even infuse empathy into basic customer-facing interactions. For example, a customer support rep can use session replay to help a  disappointed customer understand her ordering mishaps and assist with a solution.

Frustration heuristics can be used to codify and surface user pain from the raw recording stream. Frustration signals such as Rage Clicks(™), Mouse Thrashing, Form Abandonment, and Dead Clicks are then automatically surfaced as session data is analyzed, exposing revenue-impacting pain points within this digital customer journey. Signal events serve as a core building block of customer understanding, connecting to Page Insights and conversion analysis to diagnose the impact of friction on digital business and recommend action based on a fine-tuned understanding of where the customer is getting stuck and how to best help them navigate those points.

Embrace AI to Mix Evidence, Human Empathy

As businesses continue to prioritize digital transformation, the use of AI techniques brings new insight and transparency into customer frustrations and issues, as well as intelligent recommendations for how they can be fixed.  By pairing AI-driven analytics and insights with digital tools such as session replay, organizations can get a complete picture of user triumphs and struggles at the macro and micro level and make meaningful changes based on empathy and evidence to deliver a more perfect experience in the eyes of customers.