Bots are evolving quickly and companies are learning valuable lessons
A couple of weeks ago a headline grabbed my attention. The automated savings app, Digit, had come out publicly and proclaimed that it would be killing its chatbot interface, instead choosing to go with a traditional icon-drive menu. The founder even went as far as to proclaim the death of the chatbot:
“We think [chatbots] haven’t lived up to their promise,” he said. “So we are done believing they will.”
It was not too long ago that chatbots were being hailed as the next big thing in financial services. And banks rushed out to build them. Mastercard adopted KAI, Bank of America built Erica, and Commonwealth Bank launched Ceba, just to name a few. But in our experience over the last 6 months the euphoria over chatbots has waned.
New technology often carries with it a wave of excitement which overinflates expectations. It’s typically after the initial hype that the industry settles into a much a healthier perspective of emerging technology’s value. In our view, this is where Digit and the rest of the chatbot ecosystem is at the moment.
Digit realized that the chatbot interface was not the best option for its specific use case. It’s an extremely compelling fintech startup, but offers a fundamentally simple product (for the end-user). A consumer connects their primary accounts and Digit automatically calculates small savings based on cash flow which are transferred into another account held at its partner bank. Users can check their balance or cash out, initiating an ACH transfer back to the original DDA. There are not many features and functions, and having to chat with a bot as opposed to using a menu within an app would be over engineering the experience, creating unnecessary friction.
The industry is maturing into a much healthier stage of accepting the limitations of chatbot technology, understanding where they succeed and where they struggle. We think this will ultimately help institutions take smaller, more impactful steps towards successful implementations.