I rang my bank last week.
I got an answer-phone message. For the third time in two weeks, there were “unprecedented high volumes of calls.” The minimum wait time would be 20 minutes. The answer machine service suggested I use my mobile app. It would be quicker, it said.
It didn’t consider that it was the mobile app that had directed me to the phone.
I got on to the bank’s chatbot. The AI beautifully and promptly answered my first standard questions. But when we went off-piste with fewer standard questions, something changed.
I was no longer dealing with an AI system but someone typing behind the screen.
Whoever answered the off-piste questions felt pressure to respond promptly. They struggled with the correct spelling of words in their answers.
I started to giggle. Finally, the person with the poor spelling said the bank would complete my request in seven to ten days.
But the actions were never completed.
I had to drive to the nearest bank seven miles away. The helpful clerk completed the task in less than five minutes. He told me he got requests like mine every day.
This experience has made me reflect. AI undoubtedly has the potential to revolutionise how businesses operate. When implemented correctly, technology can lead to significant productivity gains, improved customer satisfaction, and even happier employees.
But technology alone is not enough to ensure productivity and customer satisfaction. The key lies in the quality of data and the people behind the systems. If they can be combined with AI, productivity gains will be significant.
Since using AI, customer service at legacy banks has gotten worse, not better. It’s as if some banks are using AI to provide the facade of modernization, without meaningfully changing their business processes.
It feels like there are several reasons for this.
The banks have adopted standardised digital systems too late. Legacy banks are playing catch-up in digitalizing their operations.
At the same time, they are suffering an identity crisis in the face of online-only banks. What is legacy banks’ market position in this new landscape? Customers find it hard to tell them apart.
AI holds immense promise for improving business operations. But too often its implementation is one inch deep.
Legacy banks have poor data management. Words are data, and if you’re feeding incorrect spellings into your AI systems, you’re compromising their effectiveness. Many companies, including my own IT firm, spend significant time sorting out poor data entries. Without clean, accurate data, even the most advanced AI systems will falter.
Lastly, legacy banks see the use of AI as an excuse to build walls between customers and the human assistance they need. This is wrong.
Technology should enhance, not replace, human interaction. Companies need to invest not just in AI and digital systems but also in training their staff and improving their data quality. In light of AI, they need to change how they run their businesses.
Here’s five key takeaways for any banks looking to implement AI the right way.
- Augment, don’t replace. Use AI to enhance human capabilities. Don’t use it to create barriers to human interaction.
- Focus on customer needs. AI implementations should solve customer problems rather than create new ones.
- Maintain data quality. Poor data input will severely hamper AI effectiveness.
- Rethink processes. Don’t overlay AI on poor existing processes. Re-imagine how your services can be delivered better than ever before.
- Stay human. The most effective AI implementations involve collaboration between humans and machines.
AI holds immense promise for improving business operations. But too often its implementation is one inch deep.
AI means banks must change the way they run their business processes. That requires good data, good people, and putting customer needs first.
The million-dollar question, of course, is will customers like myself continue to tolerate subpar service? Or will we vote with our feet and change to providers who understand this crucial balance?
I think we all know the answer.
