Back to the Future: The Rise of Voice Tech
Brands can also benefit from repeat-purchase by leveraging the technology to initiate an order in a smart way.
Voice technology is on the rise, and has been for a while. Machine learning has come hand in hand with voice recognition and AI, to create the ultimate robot friend. But where is this all heading? Below Tim Carter, Business Director at smp, sheds some light on the future of voice tech in business and consumer technology.
Siri arrived on the iPhone 4s in 2011. Amazon launched Alexa in 2014 and a series of smart assistants, including Cortana and Google Assistant, have come along since. The more we use them, the more they learn, adjusting to speech patterns, vocabulary and personal preferences. Amazon, for example, trains their speech recognition and natural language understanding systems, based on requests received by Alexa. Yet, it could be argued the more we use them, the more they learn from us.
With voice as the trojan horse utilised by different vendors to achieving Artificial Intelligence loyalty (i.e. the smart assistant being the lock-in to the platform), the long-term win is to achieve loyalty as a result of AI ‘learning’ about their users. What is clear is that machine learning will be optimised in time to learn and predict consumer preferences and purchase habits. As a result, it will recommend the preferred brand once a consumer asks for a product in the future. Current predictions indicate that by 2020, 50% of all searches will be carried out via voice, forming an opportunity for brands to become part of the conversation in the new “voice-search-to-purchase” shopper journey.
Shopping and transactions account for a mere 1% of the most common activities undertaken through voice assistants, with information retrieval representing 40%.
But let’s take a step back in time again to the Eighties. Back to the future even. As long before its ‘post’-decessors was perhaps the original Smart Assistant: KITT. The main protagonist of the hit TV series Knight Rider, KITT (Knight Industries Two Thousand) was an indestructible black Pontiac Firebird with an Artificially Intelligent electronic computer module. The “brain” of KITT was the Knight 2000 microprocessor which formed the centre of a “self-aware” cybernetic logic module that allowed KITT to think, learn, communicate and interact, with an incredible one nanosecond access time. Essentially, KITT was an advanced supercomputer on wheels, which could talk to you and search deep in itself for answers. Perhaps the ultimate smart assistant. Something that was, at the time, the stuff of imagination.
Fast-forward to now. Shopping and transactions account for a mere 1% of the most common activities undertaken through voice assistants, with information retrieval representing 40%. This is something that brands should bear in mind when investing in voice with perhaps focus on shopper-centric product features, ratings and reviews (garnered through independent and social proof) as well as product awards which could later be complemented by paid search (with Amazon planning to introduce a Google-like paid search on Alexa).
It, of course, goes beyond search. Brands can also benefit from repeat-purchase by leveraging the technology to initiate an order in a smart way. For example, a voice-controlled TV remote realising it’s low on batteries, prompting Panasonic batteries to be added to your Amazon basket, or a smart washing machine running low on its tank of detergent and reordering your preferred brand. Retailers are testing the technology, with Argos being the UK’s first to offer a shopping service via Google Assistant. This allowed consumers to reserve products in their local store using a Google Home smart speaker.
As the technology is still in its infancy, interactions with smart assistants still come with certain problems, with poor comprehension of commands and limitations in verbal output being cited as the most common. However, the reality is that ‘hands-free’ usability largely outweighs the annoyance of poor usability. And as for the inevitable future challenges smart assistants may be posed with; What if there is no answer? What if there is no stock? It’ll probably (machine) learn one over time…