Which customer bought what and when?
For any enterprise that analyze data that was the central question they sought to answer to understand their place in the market. But in this new era of disrupt or be disrupted, companies must begin asking themselves more powerful questions — like, “what are our customers going to do tomorrow?”
At the last full day of Partners 2017, Teradata’s Oliver Ratzesberger sat down with business leaders to discuss the future of analytics in the enterprise, highlighting key points in his new book co-authored with Northwestern University’s Mohan Sawhney, “The Sentient Enterprise.”
Predicting behaviors, not just transactions
For many businesses, adopting more advanced analytics practices in an environment that can seamlessly blend data sets means they can pivot from predicting transactions to predicting behaviors.
Grace Hwang, executive director of business intelligence and advanced analytics at Verizon Wireless, detailed how her company has leveraged scalable analytics to maintain the lowest churn rate in the telecom industry.
“If you are saving for the sake of saving, then people would be just waiting for your next promo,” she said.
To avoid that scenario, rather than focusing on short-term metrics, Verizon instead plays a longer game when assessing a customer’s journey. For instance, a customer that will likely buy up a sale right away looks like a win for the company. However, if that person then calls into a call center later to complain, then it’s clear offering that individual a sale wasn’t good for the company’s bottom line in the long term. The key to making those inferences is un-siloed data, so a company can have a transparent view into all facets of a customer journey, both transactional and behavioral.
Aligning culture to embrace behavioral analysis
At banks, however, predicting behavior isn’t just about knowing what your customers may do — it’s about knowing who is masquerading as one of your customers as well. That has driven the 146-year-old Danske Bank of the Netherlands and 200-year-old Westpac Banking Corp. of Australia to transform themselves into futuristic companies armed with AI- and machine learning-based fraud detection.
While that shift is a challenge in itself, another trial soon followed for their companies — culture. Getting people to work with and trust AI after years of using a different method can be a challenge.
“When you go through our organization — we aren’t that big, 20,000 employees — there are people who don’t understand,” said Nadeem Gulzar of Danske Bank. He said advanced analytics is the “ace in the hole,“ arming Danske Bank with the data that proves what its employees should be working on.
Dr. Leif Evensen of Westpac says companies still must educate employees so the impact of investing in AI isn’t diminished.
“The problem is, you can invest any amount of money in the technology and in analytical talent, but the binding constraint comes from the culture and capability of the organization — basic things like being able to frame and define a problem and being aware of biases of humans when they are trying to arrive at a decision,” says Dr. Leif Evensen of Westpac.
Preparing for the future
And as cultures evolve, companies must continue to evolve as well. So what should enterprises do now to prepare for potential sentience in, say, 20 years? It’s always hard to say when it comes to technology, but panelists had a few guesses.
“I’m a firm believer in AI assisting us in all shapes and forms,” said Gulzar. “It will be embedded in our lives to a degree we don’t understand right now, just like smartphones.”
However, some panelists believed the increased role of AI would change one job title in particular.
“Data scientists, 10 years from now, I think that job will be irrelevant,” said Satven Sangani, of Alation. “I can imagine Teradata 20.0 will come with, ‘Hey, Alexa!’ or ‘Oliver,’ and the data science can be done in two to three minutes. But the implementation will be hard; [then] the running of model and testing will be the important job.”
If that forecast proves true, it’s the young generation now that will interact with machines and models in a more natural way. And from his experiences at home, Ratzesberger has witnessed just how comfortable workers 20 years from now will be with intelligent agents.
Around last Christmas, Ratzesberger happened up on his young daughter, sitting in her room alone, talking to her Amazon Echo, which she uses to play music since CD players are no longer en vogue.
“She was chatting with Alexa,” he said. “She went back and forth ... and gave feedback as a 7 year old.” It left Ratzesberger with one question: ”Should I be scared, or proud?”