Voice assistants: what you need to know for marketing and public relations
A primer about Google Assistant and Amazon Alexa and the opportunity for organisational communications.
There was a clear message from Google’s IO developer conference last week: the future of computing and interaction between human and machine, is voice.
Voice assistants are a nascent form of technology made possible by advances in speech recognition, high speed internet connections, and raw computing power.
Number crunchers at analyst house Canalsys predict sales of 56 million worldwide in 2018, up from 33 million in 2017. The US is the largest market accounting for almost 70% of sales.
There’s still some way for the technology to go before it becomes truly mainstream.
Accents can be challenging and syntax is a work in progress. For example, Google claims that there are more than 10,000 ways to ask about something as simple as the weather.
What are voice assistants?
Amazon Alexa on the Echo platform and Google Assistant on the Google Home platform are making in-roads in the home. Apple’s Siri and Google Assistant are available on Apple and Android devices, respectively.
Siri was launched in 2011 on Apple hardware. It’s not achieved the widespread adoption that was initially expected because barking instructions at your phone makes limited sense. It’s socially awkward and text entry works just as well.
The market for home assistants started to take off with the launch of the purpose-built Amazon Echo in 2014. Amazon has since launched a range of voice products incorporating the Alexa platform including the Echo Dot, the Echo Spot and the Echo Screen.
Google followed Amazon with the Google Assistant platform for the Google Home, Google Home Mini and Android in 2016.
Both Amazon and Google are taking aim at third party markets such as automotive via licensing agreements with manufacturers.
Branded voice applications
There are four opportunities for brands to use voice assistants as a form of branded or organisational communication.
Voice assistants can be used to serve media much like any other computing device, or indeed a radio or television. Therein lies an advertising opportunity.
Second, voice assistants can be used to serve news. Both Amazon and Google serve content from more than 50 sources. It’s accessible in a bite-sized flash format, and long form such a podcasts and radio shows.
Third, organisations can build voice apps or skills. There are voice directed applications that typically pull information from the internet, and serve it in an audio format. Amazon and Google have both publish libraries.
Finally, there’s the retail opportunity. Voice assistants bring a further level of disintermediation to the web. Alexa is an audio shop front for Amazon. Google serves products and services based on voice queries, location and your social graph.
Revisiting the Turing test: challenges for public relations
A critical aspect of voice technology is that it needs to be natural. It needs to replicate human speech in so far as is possible.
In 1950 British mathematician Alan Turing developed a test to determine whether a machine was able to exhibit human behaviour.
The so-called Turing test challenged a human being to determine whether they were talking to a human or a machine during an experiment. The test is arguably within the reach of being solved by voice automation in the home.
Google may have been slow off the mark to develop a voice assistant compared with Amazon and Apple, but it’s making up for lost time.
At the Google developer conference last week it said that it would be bringing new voices to the platform. It has announced six new voices with more to come based on well-known people such as John Legend.
Its WaveNet technology from DeepMind means that it’s possible to train Google Assistant to learn new voices without thousands of hour of studio time. Fake audio will almost certainly become an ethical concern for communicators.
There’s more: a new technology called Google Duplex enables the Assistant to understand complex sentences, fast speech, pauses and long remarks.
Google demonstrated Google Assistant making calls to book appointments at a hairdressers and dinner at a restaurant. In each instance it appeared to interact naturally with the person on the end of the telephone and it was hard to distinguish human from machine. Disclosure will also become a concern for communicators.
The future is already here, but to misquote science fiction writer William Gibson it’s just not evenly distributed. We’ll have to wait until the summer for Google Duplex.