Christian Wiklund is the cofounder and CEO at unitQ.
The community bank where a close friend handles his personal finances recently called him about suspected fraudulent activity. The bank said a recent charge that their security algorithm flagged seemed suspicious. My friend told me he was extremely concerned initially. But, as it turned out, he had indeed authorized the $120 purchase—an annual recurring charge for a warehouse shopping membership.
Because he’s a data scientist, my longtime friend suggested to the person who called that the bank’s security algorithm might need adjusting or training to stop it from flagging these types of recurring charges. The person from the bank’s security department responded that their job was to resolve flagged charges on customer accounts. They said to contact another department in the bank to raise the issue.
As my friend was relaying this story to me at a recent family gathering, I instantly realized this was a tragic customer experience failure on so many levels. The episode forced me to understand that a failure in one department can have ripple effects across an entire organization, resulting in a broken customer experience.
Consider this: the bank’s failure to communicate feedback across departments made my close friend feel like his opinion wasn’t valued, which pushed him to consider banking alternatives. And equally important, it was a blown opportunity: The bank’s security department will likely continue spending unneeded time dealing with these same and similar types of expenses that are being unnecessarily flagged on customer accounts. Worst of all, this loop likely will continue unbeknownst to the team responsible for maintaining and training the algorithm—all to the detriment of the bank’s security and customer experience.
Five Steps To A Better Customer Experience Using AI
This got me thinking about the many stumbling blocks organizations are facing when it comes to customer experience issues connected to AI. Fortunately, there are multiple steps tech leaders can take to ensure AI is leveraged intelligently in the workplace to safeguard a smooth customer experience.
1. Company Education
The worker from the security department who called might not have understood the importance of training AI. That’s because this person felt there was no need to pass along to the team responsible for the security algorithm that the AI might be flagging purchases unnecessarily. Hence, to counter this, company education sessions are in order to provide employees from all functions with enough information to demystify AI and for personnel to learn how AI can enhance productivity as well as the customer experience.
2. Customer Appreciation
Speaking of education, had the bank worker who made the phone call been trained about AI, their playbook might have also required that they apologize for the inconvenience and the distress it initially caused. The bank worker, for example, might have said something different—that their security algorithms are always being perfected and that they appreciated the help in improving bank security.
3. Contextual Training
Algorithms should understand context. There was historical data—years’ worth—in my friend’s banking ledger in which he had automatically made the recurring warehouse membership purchase at the same time each year. There shouldn’t have been a fraud alert triggered in the first place.
4. Cross-Functional Collaboration
Organizations shouldn’t rely on their customers to ensure that AI algorithms are trained properly. But that’s exactly what happened here. There should have been a process in place that requires personnel, from any department, to alert the right teams when AI is wrong, when it’s burdening customers and when it’s creating unnecessary work.
5. Let Humans Be Human
Empower human decision making so that workers aren’t required to blindly follow a flag from an algorithm. Instead, let human intuition play a role. When the bank called, the person on the phone said they had no choice but to call. The call was required when their algorithm flagged a transaction. Clearly, adding human choice into this equation, and in many others, might be difficult to put into practice. However, had the bank taken the steps I’ve outlined, the $120 transaction likely wouldn’t have been flagged in the first place. A call from the security department wouldn’t have been placed, and the bank likely wouldn’t be losing a customer.
The AI-Customer Experience Crossroads
AI is becoming ubiquitous, and it obviously holds tremendous power to enhance the customer experience across all touchpoints of the customer journey. But, organizations must adopt AI wisely and guide it properly so that it empowers our workforce and serves our customers well.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
Read the full article here
