Is the Next Generation of Chatbots Here?

We are on the cusp of big changes in customer self-service.

Back in August I met with Erik Krapf, general manager of Enterprise Connect, to pitch a presentation I wanted to do at the show at the end of this month. I was explaining to him that the topic I wanted to discuss was how generative artificial intelligence is about to change everything for conversational AI (chatbots and intelligent virtual assistants). Here is what I told him:

Generative AI is giong to change what chatbots can do in a way that will fundamentally change the contact center. There is an order-of-magnitude leap in the capabilities of a chatbot or IVA that is possible when generative AI is used. Customer interactions that would have taken nine months to build historically can be put together and fully tested in a couple of weeks. And the customer experience can be so much better. Generative AI's ability to converse naturally will replace the stilted conversations we have today with systems that force a single threaded conversation to get one very specific thing done.

Back in August, it was already clear that conversational AI vendors were in an arms race to retool their products to embrace generative AI. Even though most contact center managers were not interested in generative AI for any purpose at that time, vendors saw what was coming.

Erik asked me an interesting question about this. He wanted to know what I thought it would take to get companies to start using generative AI for customer self-service?

I'd been ruminating on this for a bit, and I had a ready answer for him. My answer was that someone with a great marketing and PR department would build a jaw-drop application that was undeniably something new. An application that provided powerful and game-changing capabilities for customers with significant savings for the company through increased automation. This company's marketing engine would make a lot of noise about how cool their deployment was, and other companies would have to take notice and investigate a new chatbot.

I also predicted that it would happen before my presentation at Enterprise Connect later this month.

The other point I made was that there could be bad stories as well—embarrassing flubs that truly sullied companies' names. With any other technology I would have said that the flubs would set the industry back five years, but the power here is too big; a setback would be no more than three to six months.

We've had our share of very public flubs, including the following:

  • An auto dealership that released a chatbot that sold cars for a dollar.
  • The delivery company whose chatbot went rogue and bad-mouthed the company.
  • An airline's attempt to disown the actions of its chatbot was struck down in court.

I can't say how much these incidents might have slowed things down, I've not seen a dip in interest in companies coming to me to talk about better self-service.

But what about our jaw drop? Maybe it's Klarna just in time for my presentation.

Klarna went public recently saying that its new chatbot was so efficient that it could automate the work of 700 agents in the first month. Ugh, no exciting story about an other-worldly new capability here, just a good solid FAQ that could have been built without generative AI but would have required much more cost and pain to take to market. This was the same story that built interactive voice response jail in the 1990s. We can automate a lot of stuff; it will save a lot of money because computers are so much cheaper than people. It's a powerful story, and it could easily garner a lot of attention and lead to a lot of other companies looking for significant savings in customer service.

This is not what I wanted; I wanted a story about other-worldly customer experiences that also saved the company a lot of money. The generative AI genie is out of the bottle, the first of my three wishes is spent, and it's not going where I wanted.

Be careful what you wish for, you might just get it!


Max Ball is a principal analyst at Forrester Research.