Back in 2019, Lush opted to close its social media accounts after being “tired of fighting the algorithms”.
The cosmetics brand now engages with customers over messaging channels such as Apple Business Chat instead.
And now AirAsia has taken a similar step - closing all of its call centers. The airline has invested heavily in messaging and in particular its AI chatbot, AVA.
"...we've been able to do something really controversial … which is close our voice call centre, because what we've learned is nobody wants to sit on hold to a voice call center anymore. Life is about messaging." - Chief Customer Happiness Officer, Adam Geneave
We’ve been working with the leading low-cost-carrier in Mexico, Volaris, to help them streamline their customer support offering. Messaging accounts for most of their inbound traffic and they were able to turn off live chat in 2018. You can learn more about our journey with Volaris here.
When you make a change as big as shutting down your call center you need to be bold. Plan ahead and, when it is the right time, pull the bandaid.
2.6 billion parameters
40 billion words
341 GB of text
And the result? Cow jokes, obviously.
To be serious for a second, this is Meena - Google’s attempt at making a true conversational AI.
To prove Meena’s effectiveness, the good people at Google have also created a new metric - Sensibleness and Specificity Average (SSA). Essentially, SSA is the rate at which a chatbot provides a suitable response that’s understood by the user. Humans apparently have an SSA of 86%, while Meena isn’t far behind with 79%!
But what does this mean for brands utilizing chatbots? For now, not much will change. Brand to consumer chatbots are predominantly geared towards completing specific tasks. That means they draw upon existing customer data and Natural Language Processing (NLP) when required to, but, for the most part, they have a defined back-and-forth, Q&A structure with a consumer.
It’ll be interesting to see how Meena’s open-ended approach to conversational dialogue develops. Our current approach is to let bots handle the tasks that they’re good at while humans deal with any interactions that call for empathetic understanding.
Perhaps in the future, the process-specific bot will pass a customer query on to another bot which has more sophisticated conversational AI...
It turns out 2019 was a huge year in CX for our clients. The growth of private messaging channels was huge!
It’s taken us a few weeks to crunch the numbers, but here are some of the headline stats: