How to design a powerful conversational interface

With advancements in conversational interfaces, like Amazon Alexa and Google Home, and AI and NLP powered chatbots rapidly maturing, how can brands create a good conversational interfaces for consumers?

Kraftwerk, Linked by Isobar Co-founder, Boris Nihom, explains how their Conversational Canvas tool can help.

Create better chatbots and digital assistants with the Kraftwerk Conversational Canvas

The world of chatbots, digital assistants and smart speakers like Google Home and Amazon Alexa is rapidly maturing. More and more brands are successfully using the potential of these platforms. The number of applications is growing exponentially. And the user experience is getting better. But how do you design a good conversational interface? Kraftwerk, Linked by Isobar introduces The Conversational Canvas: a free template that helps design a digital assistant that really adds value.

Conversational Interfaces are advancing

According to recent research from Gartner as many as 30% of Internet traffic will take place in 2020 without a screen. In the US more than 10 million Amazon Alexa devices were sold during the recent holiday season alone. Google and Amazon rapidly role out their Google Assistant and Amazon Alexa technology around the world, followed by Microsoft Cortana and Apple Siri/HomePod. Meanwhile, the audience already got used to chat bots for several years via Facebook Messenger, Whatsapp and WeChat in China.

More than a gimmick: the new battle for “Share of Voice”

The first conversational interfaces performed simple tasks: a question to customer service, a simple purchase or switching on the lights at home. But today, much more is possible. Using AI and NLP (Natural Language Processing) conversational interfaces are increasingly able to deal with natural language, as we’ve seen during the latest Google I/O demo. And new hardware can work with visuals or even recognize images. A number of progressive brands are already experimenting with more complex user experiences and as a result take a significant knowledge leap. And as is currently the case with websites and apps, people will soon divide their time and attention on just a few assistants: The new battle for “Share of Voice”.

Conversational Canvas

However, a conversation interface is not a goal in itself. “We need a chatbot … something with Alexa” is never a good briefing to your team or creative agency. What is the relevance for the user? How does the proposition contribute to business goals and brand strategy? What space in the customer journey do you occupy? The time has come to take this development seriously and to develop conversational interfaces that really add value. To help with this, we have designed the Conversational Canvas. In four steps, this canvas will help to come up with better propositions for a digital assistant.

Step 1: Looking for user value, business value and brand value

The top of the canvas is divided into three sections: ‘user’, ‘business’ and ‘brand’. What user need do we solve? What is the business goal? And how does it fit with your brand?

First, consider for whom you’re going to make an assistant. Which task is he or she trying to accomplish? And in what context does it occur? When we designed The Bar for Diageo, we discovered that people who make cocktails at home are often insecure because they lack knowledge, tools and recipes. The Bar answers this need, with 75 cocktail recipes and extensive tips and tricks. Perfect for a conversational interface, especially when you have a cocktail shaker in your hands.

Then, look for commercial value. To Diageo this was obvious: more liquor sales via e-commerce. Sometimes, however, the commercial value is not per se additional sales, but a reduction of operating costs. With Volkswagen, we strive to use chatbots to create self-checking and self-service propositions for maintenance at the dealer. This helps to reduce the peak load on staff and waiting times for customers

Thirdly, a personality assistant must match the brand. For example in the Netherlands, KLM’s BB has a different tone-of-voice than the Transavia chatbot. And Isobar Australia’s Assassin Training chatbot for a famous video game is truly next level in terms of personality. Think about the tone of voice of your conversational interface the same way as you think about using brand assets on your app or website, contributing to the visibility and distinctiveness of your proposition.

Step 2: Create a unique value proposition

The largest part of the canvas is reserved to write up your value proposition. What is the unique benefit your conversational proposition offers the market? How does this help to fill the needs and barriers of the user? It’s a simple format: only [describe your proposition] can [ describe the key features]. So for example: Only “The Bar on the Amazon Echo Show” can “Help you find your perfect cocktail, teach how to make it and order the ingredients directly”.

Step 3: Choose your platform

Think on which platform your assistant will live. Will it be a chatbot? Or a voice assistant? Will it be text only, or a combination of text and image? Make sure to think again from the perspective of consumer, business and brand value. In terms of technology, it is not difficult to design for multiple platforms; DialogFlow platform can help in this instance. The key is to think carefully about discoverability and creating the optimal user experience. Plus, consider which platform will best drive commercial value.

Step 4: Bring your assistant to ‘the next level’

Now the basic idea is done, it’s time to take the next step and bring your assistant to the next level. First, you can make you assistant smarter by adding AI and intelligent algorithms. What data can be used as input? How can the assistant become smarter over time? And how can you use the interactions people have with the assistant to learn more about your customers. Simple example: The Bar teaches Diageo exactly what cocktails are popular among users. Another example is this Albert Einstein chatbot, that gets smarter with every conversation.

The second step is implementation of e-commerce. The Diageo assistant enables people to have missing ingredients delivered through Amazon Prime. And specifically for this service, we developed complete cocktail kits that contain all the ingredients for your favorite cocktail.

Third, you can make your assistant smarter by linking it to 3rd party connected products. Think smart plugs, connected car and fitness trackers. Or you can even develop a dedicated connected product, as this impressive case for video game Destiny 2 shows.

Finally you can think about combining a conversational interface with AR or computer vision. A number of companies have already successfully integrated AR into their own app, such as IKEA and Sephora, but there are little to no cases as of yet, that make the combination with a conversational interface. The Amazon Echo Look still needs to be improved, but this fictitious mock-up for Ray-Ban sunglasses shows what is possible.

In short, armed with the Conversational Canvas, nothing stops your organization to think in a structured way about conversational interfaces. So download the canvas, gather a team of entrepreneurial colleagues, organize an internal workshop and think how clever assistants can create value for your customers, your brand and your business. And take the lead in the new battle for Share of Voice.

Boris Nihom, Co-founder Kraftwerk, Linked by Isobar