Your Genesys Blog Subscription has been confirmed!
Please add [email protected] to your safe sender list to ensure you receive the weekly blog notifications.
Subscribe to our free newsletter and get blog updates in your inbox
Don't Show This Again.
In the 1990s, flashy, neon-colored pants were the style, frizzy rock star hair was in its prime and the video for “Ice Ice Baby” was on rotation on MTV. And, in the 90s, I could also look up a person’s name, address and phone number in the white pages of my local phone book. Sounds scary—anyone could look up your information using your first and last name. But now, the most advanced artificial intelligence (AI) applications don’t even need that personally identifiable information (PII) to generate tangible business results.
By profiling, like-for-like comparisons easily can be made in a data model without identifying a single individual. Machine learning treats all data as 1s and 0s—humans coined the term PII and created regulations like HIPAA, FedRAMP, and PCI-DSS. I’m not saying that these terms and regulations aren’t significant; they’re necessary to some degree from both a legal and consumer confidence perspective. And they have inherent financial repercussions if not properly managed. But they aren’t necessary—and they don’t have a place—in the world of AI and machine learning. High-quality predictive algorithms don’t require you to identify who your customers are to produce accurate and game-changing results.
AI and Routing Get Maximum Results
Take the financial sector as an example. Tim has recently graduated from college and gotten his first full-time job. He has saved enough money to make a down payment on a home. His first stop of the home-buying journey is to hop onto the internet. Tim visit’s his bank’s website to research mortgage options. While doing so, he interacts with a chatbot and obtains some very reliable information, including answers on how to apply for a mortgage. The bot offers a call back from a mortgage specialist at a time that’s convenient for Tim. And, even though this is extremely convenient for him, Tim gets distracted and logs off. Three hours later, Tim returns his attention back to his home purchase search and decides to call the phone number the bank presented to him.
Next, Tim’s bank uses an AI solution with a model built for someone like him. After Tim dials the phone number on the mortgage page of the website, he’s routed to an agent named Deb. The AI algorithm that connected Tim to Deb has no idea who Tim is, but it does know that Tim is located Nebraska, falls within the age bracket of 20-25 years old, is a bank member, has checking and savings accounts with specific balances, and was on the website looking at a series of mortgage products but never applied.
The algorithm knows even more data elements. And not only are these data elements factored in, but the algorithm also knows the relative predictive value of each of the previous traits—and how to weight them accordingly based on historical, results-driven data. This AI solution also knows that the bank has a mortgage specialist named Deb who’s ready and available; it routes the call to Deb.
There’s another reason why the system chose Deb. The historical data shows that Deb is the best agent to maximize value for someone like Tim, calculated by all interaction data associated with the call. It not only considered Deb’s skillset and availability, but it also took into account her experience, agent group, training profile and everything the customer experience department knows about her. She has a specific DNA, or profile, that maximizes value for specific combinations of customers and interactions. This call was routed to Deb because the AI software knows, at the interaction level, that there is a 90% chance that having Deb take the call will result in the loan application being submitted.
The AI software also knew that routing the call to Dave, another licensed mortgage specialist at the bank, created a 60% chance of Tim submitting his loan application. Instead, Dave was routed to interactions and customers where he’s predicted to have the best outcome, resulting in higher revenue for the bank and happier customers. Dave’s performance will be higher because he’s receiving calls for which he can naturally maximize value. AI and machine learning uncover this precise, real-time and extremely valuable optimization.
The software combines mathematical algorithms with the power of cloud computing to improve results for the business, its customer experience center and sales. And, the public API that returned the predictive interaction score for this entire scenario didn’t know Tim’s address or phone number. If you want that, you’ll have to look in the phone book.
Learn more about predictive routing at our resource hub.
Subscribe to our free newsletter and get blog updates in your inbox.