On Thursday 6th April, I attended a very impressive voice tech session hosted by the Hoxton Mix Collective. The impressive line-up of speakers included – Oscar Meary, co-founder of Opearlo, talking about ‘Alexa Design and Best Practices’ ; Marc Paulina, interaction designer from Google, on ‘VUI Prototyping Tools, Design Sprints and Research Best Practice,’ and Dean Bryen, voice/AI evangelist from Amazon.
As it was also the day that Google Home was released in the UK, this seemed an appropriate time to take stock of the current state of voice tech and associated opportunities.
So, what then did I learn?
A major issue in this space is around high consumer expectations and the challenge of managing them. We have all seen films like Her and Ex Machina , and we therefore think we know what AI will be like in the future. As a result, a rather unrealistic benchmark has been set for the ‘real’ voice tech that is being developed today.
In addition to this, people are less tolerant of voicebots than they are of other digital assistants. This is because voice (talking) is the way that humans naturally prefer to communicate with each other as the process is, generally, an easy one. If a voicebot cannot meet this level of ease or proficiency, then the experience can be frustrating.
A clear message from Oscar Meary of Opearlo (a voice design agency) is that despite all the excitement, this is a very new area. Companies looking to develop a voicebot need to be realistic about what can be achieved successfully. One of the main problems can be around integration – an organisation’s digital design department may create a voicebot, but if this is not connected to back-end systems, then there is no sense of a joined-up user journey for the consumer.
Games or utility?
One speaker made the case for a more conservative ‘gamified’ approach, focusing on simple conversational design voicebots or gamebots. This simpler experience takes the user into a different world more easily than a utility bot, and because it has a more light-hearted approach; individuals are more likely to be forgiving of functionality issues. Content examples in this space are around games, such as pub quizzes and ‘would you rather’ games.
The current scenario with voice tech and bots reminds me of the early days of the web, when brands rushed to acquire digital real estate or shopfront ‘ brochureware’ websites; where they could show of their products. It took some time before marketeers realised that consumers were not interested in visiting websites housing uninteresting and functional product information. This mistake was often repeated during the peak period of app development, where brands rushed to create their own apps, and this could easily happen again with voicebots. As always, the solution lies in providing useful content – what value can be added and how brands can add to the user experience.
Perhaps surprisingly, the technical build of a voicebot is relatively easy via the use of available SDK’s, but the real work goes into the design of the customer interface and experience. Key areas to address in the design process include – setting expectations, limiting choices and options (to keep the process simple) and minimising pressure on the user. This last one is interesting because, again due to the nature of human voice interactions, consumers can feel under pressure if the interaction process does not feel relaxed or natural enough.
I learnt some interesting voice tech language as well, including the importance of spending time on ‘edge cases’, ‘half happy paths’, ‘utterance expansion’ and taking a heuristic approach. All these techniques and approaches are concerned with fully testing a voice interaction (such as an Alexa skill) to ensure a successful consumer experience.
NB; for the avoidance of doubt, the video below is an example of a poor automated interaction and user experience.
Voice as first brand touchpoint
Voice tech is important for brands because, in the future, voice will be the first brand touchpoint for the consumer. This is already taking place on mobile (e.g. with Siri), in the home (with Alexa and Home) and is moving into the automotive sector. As the importance of voice and voice tech grows, it is essential that an approach is effective and integrated, across different platforms.
Voice is also likely to become central to the area of identification, with the increasing use of biometrics. The human voice is considerably more distinctive and individual than finger prints, and will increasingly be used for identification purposes, as in voice PINs for example.
Most exciting for brands is the area of sentiment analysis. By listening to a user’s tone of voice, brands will not only be able to understand what consumers want, but also how they are feeling.
Google’s perspective is that voice technology, AI, and bots are all converging and will eventually represent one technological interface. Amazon’s ambition is to provide ‘the world’s most powerful voice service’ that will ‘power the connected home’. Dean Bryan, from Amazon, sees Alexa having an impact with audio, in the home, in car and with partner brands, for example Uber and Just Eat. I rather liked the fact that Dean sees the Star Trek computer as the apogee of voice tech and the product they are trying to emulate.
The key to any successful technology is that it should be invisible to the end user. A fitting end to this summary piece, is the quote shared in the session from The Media Equation (Reeeves and Nass 1996) – ‘…Individuals’ interactions with computers, television and new media are fundamentally social and natural.’
This is especially true with voice tech. The intrinsically ‘human’ nature of the automated voice, means that people will want to interact in a very natural and social way. Brands, becoming involved in this space, will need to ensure that this happens.