Designing Services Around AI
The nature of web design is changing, with artificial intelligence opening up new possibilities for customer conversations.
With the likes of Alexa, Siri, and chatbots in mind, there’s an increasing focus on software and user interfaces which actively take the reins and guide the user through online resources, rather than the user pointing and probing their own way through the pages of traditional websites.
The fact that this style of interface relies on artificial intelligence has brought the possibilities presented by AI back into the media’s attention. We’ve lived with different brands of AI for so long, that we’ve become slightly oblivious to its presence, and it’s only when AI takes on the guise of Siri or a chatbot that we become reacquainted.
Chatbots and virtual assistants are just a different face of AI and as much as they tap into our sense of science fiction, their practicalities are clear.
On the one hand chatbots and virtual assistants can take the customer conversation to places where traditional websites and apps can’t, seeking out and interacting with customers on the likes of Facebook Messenger or in the home; on the other hand, they offer a practical way of making complex services much more accessible and providing huge new levels of efficiency. By 2022 it’s predicted that banking and healthcare can save up to $8billion in efficiencies.
Public services could also stand to benefit and cash-strapped local councils seem eager to realise these efficiencies when redesigning their online resources on cloud based platforms.
The Conversational UI
Enfield Council standout as a first mover in terms of webchat, having recently launched Amelia - a “cognitive virtual agent” which will help alleviate the burden of 55,000 phone calls a month and 35,000 website visits a day.
The queries directed at chatbots are fairly formulaic due to the fact that the service and the user have probably been brought together by common intents. Despite a fairly strong starting point of mutual understanding, council services have a broad user base, so the ability to process different styles of questioning is vital.
This makes things complicated and websites based on more traditional information architectures with point and click workflows will continue to be a sensible option for many services.
Nevertheless, a good chatbot can incorporate machine learning and go well beyond the rigid scripted answers which are merely automated FAQ’s. A good conversational UI can learn, apply context and guide the user with much more clarity than the guesswork it might take to decipher a complicated council website, loaded with individual services.
The Screenless Interface
With a slightly different take on artificially intelligent software, Aylesbury Vale District Council have plans to go down the screenless route, creating an Alexa skill for customer services - a plan which showcases technological vision but as importantly, it illustrates how more and more councils are devising impressive long term data strategies and cloud enablement policies via the likes of AWS.
Screenless interfaces like Alexa take the challenge away from the graphical user interface entirely. Personal assistants such as Siri and Alexa are becoming particularly impressive because the queries they answer can be loaded with context.
This could deliver huge advancements in terms of accessibility - those with the type of disabilities which mean that traditional web resources can be a struggle, could prompt information from personal assistants in a similar manner to having a council representative right there with them.
Future Alexa integrations using conversational APIs will give third parties a much greater ability to integrate personal assistants with their own diverse product offerings.
UX as the Driving Force
With regards screenless interfaces like Alexa, customer journey maps become more complicated as the touch points the user is unconsciously interacting with become more convoluted. Good Service Design is essential both in ascertaining whether a project is suited to this kind of development, and in making it work.
Bots and virtual assistants need constant iteration and feedback loops to refine what it is they’re trying to achieve and for AI based products to work, usability testing and the UX profession in general will be a vital partner.