How Many Chatbots Does it Take to Book an Uber?
One of the reasons doubters gave for assuming Uber wouldn’t work, was that ride-sharing essentially performs the same service as taxis.
But after people book with Uber once, it seems crazy to do it any other way. Not having to deal with a person when booking is a big part of the appeal. Taxi companies have since been forced to play catch up and offer apps that allow you to book in a similar way.
“By reducing the consideration of social nuances, the cognitive effort involved in the interaction is reduced. This favours our psychological preference for simplification.”*
Given the choice, many of us would reserve our cognitive efforts for things other than booking a ride.
Artificial Intelligence and chatbots have the potential to offer streamlining like this to the customers of a diverse range of Australian businesses. A chatbot doesn’t eliminate the need for the customer to use the business (or staff in many cases), it just simplifies the entry point.
“Bots, like the apps before them, aren’t a panacea. Bots won’t save your business or make you great if you don’t already care deeply about your customers. But bots can make it easier to interact with people in contexts that feel safe and familiar. If brands and companies seize the opportunity to meet people where they are in their own technological evolution, they’ll be rewarded with relationships that endure.” Chris Messina.
DiUS recently had the opportunity to bring chatbot nibby to life for leading Australian health insurer nib health funds (nib). Nibby lives on the nib website and its primary function is directing people to the right area of service before they speak to a consultant.
The main objective for nibby was reducing the amount of time wasted by contact centre staff on queries they couldn’t address. For example, international students need a specialised consultant, so it’s beneficial if they connect directly to that team. The secondary goal was to allow customers to self-fulfill where appropriate.
The decision was made to create the chatbot with Amazon Lex — the AWS service for building conversational interfaces for any applications using voice and text — as it would fulfill nib’s requirements. After discussions and a two-week proof of concept, the project went ahead.
Nibby went live four weeks later, with the project complete and handed over at ten weeks.
“Thanks to our partnership (with DiUS) we’ve been able to build nibby in a faster, more agile way which has reduced the operational complexity and the cost of running and managing the chatbot.” Brendan Mills, CIO, nib.
At handover, nibby was delivering 70% successful conversations and of those 26% were completed by nibby alone.
Mauricio Perez, DiUS Senior UX Consultant, puts this success down to team collaboration between the tech team, chat support staff at nib and DiUS.
“Even if a conversation is deemed successful, or fully completed by nibby, that doesn’t tell us what the user experience was like, so we tested this regularly from day one.”
Daryl Wilding-McBride, Head of Tech Horizons at DiUS, has been closely following developments in AI since completing his Master of Engineering Science focusing on Artificial Neural Networks and Speech Recognition in 1992:
“Back then we were just trying to get voice recognition to work. These days we get annoyed if Siri misses the odd word.”
Daryl (who is now completing his PHD in Deep Learning and Medical Biology) explained that recent advances in AI have been powered by Deep Learning, the evolution of those early neural networks. He said:
“Deep Learning is now state-of-the-art in perceptual tasks such as image understanding and natural language processing. These are key elements required for systems to take a significant leap in being proactively of service to the user. This change is already happening; as it progresses and moves into the mainstream people will see systems as valuable collaborators rather than reactive servants.”
He sees opportunities for Australian retailers to use AI to personalise the physical retail experience — much like what is currently experienced in online shopping. You could opt in to be recognised as a regular customer, have previous purchases inform your next visit and be able to compare all relevant products in one place.
But what really excites him is the potential to free up some space in our brains:
“If you think about the computing systems that we have been building as an industry over the last 30 years or so, they have become faster and prettier, but the cognitive burden on how to use them still rests with the user. With AI, that cognitive burden shifts from user to the system. These advances are going to mean that the system can perceive how it could be of service to you.”
DiUS is looking forward to supporting many more Australian businesses through their AI journey. Having grown our machine learning expertise since 2004, we are excited about the progression of technology that will underpin our big ideas.