Customer Journeys, Big Data and Astronomical Opportunities

How AI Is Changing the Universe of Customer Experience

Exploring and eventually, managing customer journeys has several similarities to astronomy. Customers are like stars, and their journeys are like planets gravitating around them. There are billions of planets living across time and space; they follow rules and patterns, but they are all unique. And much like with astronomy, customer experience professionals can apply artificial intelligence (AI) to assist with the science of customer journey management.

The 2017 article, “Artificial Intelligence: NASA Data Used to Discover Eighth Planet Circling Distant Star,” discusses how NASA trained computers to identify exoplanets in the light readings recorded by the Kepler satellite—the minuscule change in brightness captured when a planet passed in front of or transited a star. The AI algorithm sifted through astronomical volumes of data and found weak transit signals from a previously missed eighth exoplanet orbiting Kepler-90 in the constellation Draco.

“Machine learning really shines in situations where there is so much data that humans can’t search it for themselves.” – NASA

AI holds immense value in helping humans solve complex problems. And in the same way that NASA used AI to collect volumes of data and rules about exoplanets, customer experience professionals can use the technology to understand customers—observing them and establishing simple and general rules about their behaviors.

“This finding shows that our data will be a treasure trove available to innovative researchers for years to come.” – NASA

The Big Bang of Customer Journey Management

Customer journey management has its roots in the telecom industry; it started with the basic concepts of ACD. While this technology was effective in distributing a large volume of incoming calls to a selected group of people, it had major gaps for customer experience—voice-centric, all calls equal (FIFO), no personalization and poor reporting. The only way to operate an ACD was to adequately staff the call center to prevent capacity overflow. It is entirely left to the agent to recognize, understand and know the customer and his or her journey.

The geocentric model of the antiquity, an ACD put the observer at the center of the universe, the customer journeys all seen as equivalent as it was impossible to account for their unique characteristics.

The next generation of contact distribution technology was an improvement; it introduced queue-based routing. Using rudimentary qualification solutions (1-800 numbers and IVR), it was possible to differentiate the intent of the customers and, therefore, serve them in a different way. Contact center operations can staff the agents per channel and queue using service-level reporting. Although queue-based systems improved efficiency, they still fell short of helping the agent recognize, understand and know customers and their journeys.

The heliocentric model of Copernicus and Galilei puts the sun at the center of the universe and describes the planets solely through their orbit. In a similar way, a queue based system assumes that all customer journeys all into static predefined categories.

The third generation of interaction routing technology introduced context and skills-based routing. This breakthrough brought with it the ability to distribute interactions based on virtually any data available in the systems—across any channel. Routing rules can use CRM data, omnichannel journey data, agent data and more to decide how to match a customer interaction and a skilled agent. This abundance of data and flexibility opens the opportunity for organizations to personalize—at will—every customer experience. And customer experience professionals could focus on modeling the business rules and agent skills, and then optimize them using analytics.

Like modern astronomy of the 20th century, the abundance of increasingly powerful instruments, a vast community of scientists and diverse experiments allowed science to make strides in understanding the solar system, planets, various stars, galaxies and the entire universe. Scientists now focus on developing new models and proving them with the observational data they collect.

Many wonder if that’s it—if interaction routing has reached its ultimate maturity and if customer experience professionals now can operate their systems in the most optimal ways.

The answer is clearly no, and the reason is inherent to the problem we’re trying to solve: Every customer journey is chaotic and unpredictable by nature, and every attempt to personalize them with human-defined rules quickly reaches practical limits. According to Gartner, “What’s important to remember is that customer journeys aren’t created; they’re discovered. When we try to create journeys, we fall into one of these two traps: we either hallucinate customer needs or throw away the customer experience playbook altogether and focus on the needs we know intimately: our own.”

This challenge is a consequence of the exponential growth of digital touchpoints, through the web, mobile, contact center, back offices and branch offices. Consulting firm McKinsey found that 56% of all customer interactions occur during a multi-event, multichannel journey, while 38% of all customer journeys involve more than one channel of interaction. Combined with the variety of customer profiles, intents and agent profiles, there is virtually an infinite combination of factors that influence desired business outcomes.

Astronomers are constantly testing their models against new data; they are frequently compelled to revise their understanding of nearby or remote stars and planets because those models cannot properly describe the observations. This process of continuous learning is inherent to every practice, including customer experience.

Rethinking the Customer Experience Universe

Customer experience professionals realize that the complexity of the customer journeys cannot be managed with enough efficiency and effectiveness using rule-based systems. And alike astronomers, they need to constantly rethink their understanding of the customer experience universe.

Skills-based routing systems have these fundamental limitations:

  • Distribution algorithms are design to optimize service levels— an average speed of answer, agent occupancy—not business outcomes like first contact resolution, Net Promoter Score or retention.
  • Rules configuration is based on assumptions, such as the relationship between business outcomes and service levels or the trade-off between priority, skills and wait tolerance.
  • Rules account for a limited number of customer or agent attributes and can omit some that have a strong influence on business outcomes.
  • Agent skills referential and skills levels are loosely related to business outcomes.
  • Rules are not robust enough to adapt to changing conditions.
  • Rules and their underlying assumptions are difficult to validate using analytics data, such as what-if scenarios.

According to Forrester, 73% of data is left unused by companies; there is a goldmine of value in data just waiting to be put into action.

Astronomers are confronted with an abundance of data and the challenge to conceive new models and validate them. Lots of scientific data is unexploited by lack of human brain-power availability and discovery opportunities are missed.

The Next Step in Customer Journey Management

Predictive routing is the newest generation of customer journey management technology. Powered by AI capabilities, predictive routing uses historical performance data and matches customer and employee attributes to predict which contact center resource is the most likely to achieve targeted business goals. The routing engine enables organizations to deliver the sales, marketing and service outcomes they want, including higher customer satisfaction, increased employee efficiency, decreased costs, better collections and revenue, reduced handle times, and improved FCR. Genesys has experienced numerous successful deployments with proven improvements, for example:

  • More than 4 points improvement in NPS at an Asian-Pacific telco
  • Over 3% FCR at a European outsourcer
  • Over 3% increase in retention and 7% drop in average handle time at a Northern American telco

While queue- and skill-based routing rely on static decision-tree logic and pre-set criteria, predictive routing uses historical and real-time data, along with AI. It continuously and automatically uncovers meaningful factors that influence the outcome of interactions between customers and employees. Through AI, predictive routing builds models from aggregated customer profiles based on a favored communication channel, products purchased, past service requests, recent transaction activities, etc. This combines with employee profiles, such as tenure, knowledge, skills, interaction history and business outcome data to predict the optimal customer-employee match for the desired result. Customer and employee data models are continuously updated to improve the experience with every subsequent interaction.

As a consequence, contact center operations are shifting their focus from maintaining manually increasingly complex routing rules and skills. They’re letting AI dynamically build the predictive models that drive a business outcome. This frees up contact centers to focus on big data sources that feed machine learning and improve meaningful business drivers that AI uncovers in the data.

With the Kepler space telescope, NASA lets AI do the heavy lifting to discover signals from the data that they gather from the universe. Likewise, AI is changing the universe of customer experience. The treasure trove of customer journey data is an astronomical opportunity you should not miss.

Learn more about leveraging the power of AI to turn your data into actionable gold and connect your customers to agents that will drive results. Check out the Genesys on-demand webinar, Technology Insights: How AI is Driving a New Era in Customer Engagement.

Find out how to use predictive routing and AI to engage the right customer at the right time.