Will conversational interfaces sound the death knell for websites?

Websites are dead! Again… At least that's the case according to Alex Spinelli from Liveperson, one of the leading companies in the live chat and conversational interface space. Spinelli espoused the virtues of conversational user interfaces (CUI) in his talk at SXSW this year and proposed the rather controversial hashtag #RIPWebsites. 

So, are websites really going to die? Or is this another case of a new channel (or in this case it's really an 'enhanced' channel rather than new, but we will get to that) that consumers, members, and citizens can optionally engage with? I recall Java apps, Flash-based Rich Internet Applications (RIAs), mobile apps, and social media all threatening to kill off websites. Could messaging and conversational interfaces really see to the end of websites, or is it all just talk? 😜 Let's find out. 

What is a conversational interface? 

A conversational interface is an interface that allows for a human-to-human like two-way interaction between a user and a system, either by way of voice or text. Put simply, it’s an interface that allows a user to write/type, or talk, in natural language. 

This incorporates all types of messaging and conversational interfaces, from the apps you use every day like Facebook Messenger, WhatsApp, Viber, iMessage etc, to live chat, chatbots, voice-driven digital assistants like Google Home, Amazon Alexa, and Apple Siri, as well as onsite or in-app messaging like we see with Airbnb and LinkedIn. 

Voice or text?

Either and both. A conversational interface can be completely voice-based (e.g. an interaction with a digital assistant like Google Home), completely text-based (e.g. an interaction with a bot via Facebook Messenger) or a combination of both (e.g. voice input into Facebook Messenger with responses provided in text or multimedia). 

Why is conversational design important? 

Here’s a scary statistic for you: by 2020, the average person will have more conversations with bots than with their spouse (Gartner). Chatbots, digital assistants, and messaging are now big business. 

The global chatbot market is expected to reach $1.23 billion by 2025 with a compounding annual growth rate of 24.3% (Grand View Research). It is predicted that chatbots will be responsible for cost savings of over $8 billion annually by 2022, up from $20 million in 2017 (Juniper Research). 

Digital assistants are now commonplace. If you haven’t been gifted one recently, you’re the odd one out. In 2018, the smart speaker market boomed, with 78 million units shipped, an increase of 125% on 2017’s figure (Canalys). 

According to Statista, as of January 2019, the top three most popular messaging services in the world, WhatsApp, Facebook Messenger, and WeChat, have all pushed past the billion mark for monthly active users. While this clearly shows that more and more people are signing up to these services, the stats also show we are using messaging more and more. In the US alone, time spent on messaging apps has doubled from six minutes per day in 2015 to 12 minutes per day in 2019. 

The Chinese market in particular has taken to messaging in a big way with 50% of WeChat users spending 90 minutes per day inside the app (Quartz) and WeChat itself accounting for ~35% of time spent on mobiles (Business Insider).

But, messaging is personal...

Sure, that might be the main way you use messaging at the moment. And while we can all point to a terrible live chat or chatbot experience (I’ll put money on it either being with a bank or a telecommunications company), we all want to be able to interact with businesses via this channel because we’d rather not make a phone call!

In their 2018 Why messaging businesses is the new normal report, Facebook points out that on Facebook Messenger alone, people and businesses exchange 8 billion messages per month (and that figure is growing). 

“Business messaging has definitely gone global – and people's interest is only growing. Across Colombia, Germany, Mexico and the US, 71% of people surveyed say they're open to messaging businesses.”

We’re now seeing the major players focus their attention on business messaging with Facebook itself continuing to push Messenger for Business, WhatsApp launching WhatsApp Business last year, and Apple Business Chat now in Beta. 

It's only a matter of time before this will become an important channel for all brands in Australia (if it isn't already). 

Humans or robots? The role of artificial intelligence in conversational interfaces 

Many chatbots and messaging bots are scripted, handcrafted, and query-based. That is, the author/builder has specifically scripted the available questions and corresponding answers and expects a certain level of linearity to the user's journey. This is extremely limiting and requires an incredibly extensive decision tree and/or question and answer database for the experience not to result in a dead end. 

With the power of artificial intelligence (AI) and natural language processing (NLP) it is now possible to apply machine learning to teach bots how to undertake the many permutations of conversation paths. And you no longer need to be Microsoft, Google, Amazon, or IBM to take advantage of this technology. In fact, it is these companies, and others, that are democratising AI and machine learning by providing code-free or low-code tools for implementing AI into your own systems, and marketing technology. Microsoft has its Bot Framework, Google has Dialogflow, Amazon has Lex, and IBM has the most recognisable of them all, Watson Assistant. There are of course others, but those are a good place to start. 

Combining these frameworks with personalisation data of the conversing visitor, from your marketing automation tool or data management platform (DMP), can result in a highly tailored and human-like experience (or something completely creepy if you get it wrong!). 

For a more detailed explanation and example of how AI and machine learning can make a conversational interface more intelligent, read this article from Smashing Magazine. 

#RIPWebForms?

Perhaps rather than #RIPWebsites, conversational interfaces could lead to the death, or at least the reduction, of boring old web forms. No one likes filling out a web form but they are currently an essential part of how we use the web. Conversational interfaces can reduce the cognitive strain of completing web forms by transforming the act of requesting a user’s details into a conversation format, field by field, question by question. 

Miklos Philips provided a number of examples in his article The End of Web Forms where brands have utilised a conversational interface to replace a traditional web form. One such example was from Jack Insurance in the UK which has replaced the quoting process with a simple and elegant conversational interface. 

It is even possible now to turn your existing web forms into a conversational interface. Space10, a research and design lab in Denmark, has open-sourced a JavaScript plugin that does exactly that. 

Problems with conversational user interfaces

For every Alex Spinelli espousing the benefits of conversational interfaces, there are a bunch of people pointing to its failings. 

Let’s start with those web forms we just mentioned. While in some situations a conversational approach to obtaining a user’s details makes sense, it doesn’t work well all the time. Shekman Tang, Product Designer at Intercom, described this problem eloquently (and from experience) in his article The cult of conversational design: why forms aren’t dead yet. As a business based almost entirely on messaging, this is rather compelling. In short, its user testing exposed a few major issues: 

1. Each field in the form was delivered as a full sentence, forcing the user to parse each sentence individually (slowing them down). 
2. There was no obvious way to skip a question and the request for the information felt mandatory (causing confusion).
3. Overall, users provided the feedback that it felt ‘spammy’, 

The next cautionary tale regarding conversational interfaces is from Alex Mohebbi, who described why conversational interfaces are taking us back to the Dark Ages of usability. He proposed three key problems inherent in conversational interfaces: 

1. Low discoverability and understanding of system scope: CUIs lack Discoverability and Understanding, two key characteristics of good design as described by Don Norman in his popular book ‘How to design everyday things’. This means that users don’t know upfront what’s possible with the interface. They need to be told, or find out through trial and error. Furthermore, Mohebbi references the Uncanny Valley theory that posits that a user’s expectations become more unrealistic as the conversation seems more human. Put simply, it sounds human, so I expect it to be able to do everything! ;)
   
2. Conversations are inefficient for complex processes: There are many actions or processes we undertake online that involve a certain degree of complexity. Think about the process of comparing and booking hotels, or comparing and purchasing a new tech gadget. The decision factors and visual outputs go beyond what is possible in a pure conversational interface. Conversation Architects and Designers (yes, those are real job titles) have attempted to solve this problem with the use of visual design patterns like cards and carousels, but these elements can’t satisfy every situation and they start to beg the question - is this a chatbot, app, or a just a website? The Conversation Design team at Google used a simple example to describe this problem - “...dialogue works well for the task of finding a restaurant’s business hours, but it feels clunky for browsing a dinner menu.”
   
3. Difficulty in articulating commands: Conversational interfaces work well when the user knows exactly what they want. When there is uncertainty or where there is an exploratory or browsing nature to the process/action, CUIs start to become difficult to navigate. This is in large part due to the uncertainty about what command to articulate. In essence, these are processes or actions we would never undertake in a real conversation, and therefore a conversational interface doesn’t make sense. 

Another compelling critique of this user interface pattern is from the renowned Nielsen Norman Group. Its recent usability research on Intelligent Assistants using CUI mentions that “observing users struggle with the AI interfaces felt like a return to the dark ages of the 1970s: the need to memorise cryptic commands, oppressive modes, confusing content, inflexible interactions – basically an unpleasant user experience.” Ouch. 😬

But we got through the 1970s (and many more decades), and I’m confident that conversational interfaces will get through this developmental period too. It’s still early days for this user interface pattern. 

I still want one! 

As always, start with your customers (or members or citizens). Why would they use a conversational interface with your business or organisation? What kinds of questions would they ask? What problems can you solve through this channel? How can you ensure the experience feels as human as possible? What should you leave out?

Google summarises it nicely:

1. Start with what humans do
2. Adapt to the technical limitations (i.e. the things computers can’t do as well as humans)
3. Leverage technical strengths (i.e. the things computers do better than us) 

Then consider how deeply you want to dive in. You could start small by adding a simple live chat tool on your website, and open your organisation up to business messaging on Facebook Messenger and WhatsApp. Or perhaps dive in deeper and implement an AI-enhanced chatbot; or go all-in with a platform-driven, custom-designed conversational interface from the likes of Liveperson or Intercom. 

So, are websites dead? 

No, the funeral has been put on hold, again. I’d recommend a strong dose of skepticism whenever anyone calls out the death of anything. It turns out that Spinelli’s boss at Liveperson, CEO Robert LoCascio, had a crack at #RIPWebsites on Techcrunch at the start of 2018, going as far as boasting: "I am going to make a bold prediction: In 2018, we will see the first major brand shut down its website." It's 2019 and I haven't heard of any major brand shutting down its website… 

Conversational interfaces are like social media and apps - they are an important channel to consider investing in (if you haven't already) and they should complement your other channels, including your website. We would therefore encourage all marketers and digital professionals to put conversational interfaces (and the underlying technologies described above) on the agenda for consideration for your digital roadmaps in 2019 and beyond.

Give me more! 

Keen on this conversational design thing? We can help of course! We’d love to chat with you… 😉 

I recommend taking a look at Erika Hall's book Conversational Design. Also have a read through Google's extensive documentation on Conversation Design. 

Happy conversations!