I wanted to give some insight into what we are working on, how nib is working towards the 5 levels of Digital Assistants. Originally coined by Rasa and a great way businesses can align their roadmaps to true conversational AI.
In 2 years, we have deployed Digital Assistants in the following places: Webchat, Google Home, Amazon Alexa and most recently Telephony. It is still early days as these chatbots only handle simple question and answer pairs. Although, Google Duplex showed the world what the next level of Digital Assistants can do for us every day. Transforming to opensource conversational software allows our Digital Assistants to handle contextual conversations.
At nib, this is the future of Digital Assistants at each level:
Through this article, I will prove examples relevant to Australian Healthcare. So how do Digital Assistants help when it comes to Healthcare?
This level of assistance is as simple as receiving an SMS or push notification from the nib app. A way for us to send updates to our members.
Example: You are a member of nib and you have signed up to online services. As we reach the end of the financial year, nib prompts you to log in to online services and download to your tax statement to assist in making your tax return.
In this case, if you wanted to ask can I get my tax return from the year before, nothing would happen.
This in most cases is the type of chatbot you have encountered. This kind of chatbot allows you to ask simple questions which provides a simple FAQ response. The answer often comes in the form of a website link.
Example: With Level 2 assistants, when you ask our chatbot a question on the nib website you always get a response. But, you can’t switch context and conversations are always linear. The response always comes with general help unless something is too complex. At which point a consultant connects to the conversation.
Creating context is something that you can design for chatbots. But as designers and developers know you rarely get the response you expected. Conversations are unique and not linear so there isn’t an option to catch all and predict what the user will say. Contextual conversation means understanding every part of the conversation. Even when you are not expecting it.
Example: Level 3 assistants, allows you to get a quote depending on your own circumstances. The chatbot asks questions in relation to what you need coverage for and helps you select what is best for you. Like, do you wear glasses? How often do you visit the dentist? And whether you use services like a chiropractor, osteopath and physiotherapist.
As we know from our own relationships we get to know each other over time. The same goes for Digital Assistants. At level 4, a Digital Assistant will learn when it is best to engage you. It will remember your preferences and what works best for you. A truly personalised interface.
Example: At this level, the Digital Assistant knows you in much more detail. It doesn’t need to ask you every question. It knows what is best for you and when you can be getting more from your membership.
In the future, entire groups of Digital Assistants will understand our members and run large parts of the organisation. From personalised health assistance, marketing, sales, HR and finance. This is a vision that is achievable in a decade.
Example: For level 5, we see Digital Assistants working end to end — checking claims, processing sales and creating personalised health care. A level at which we see Digital Assistants running back end process. Unlike the past 4 examples where you can understand human involvement.
AI helps understand many data points of your membership and helps you take advantage of what is best for you. Freeing you up from making the decision.
For level 5 Digital Assistants to work we will need to see a massive shift in society. The way businesses operate and how users choose to interact with them. The leaders in this field know the importance of investing in this technology for long term gain. Failure will see these jobs outsourced for short term profit.
We are already starting to build level 3 assistants with opensource tools like Rasa. True personalisation will make decisions you don’t want to so you can get on and live your life.