My concern about voice assistants in retail

It's a sight that is familiar to any seasoned customer service assistant. You see a customer browsing with intent, leaning in to a shelf to take a better look at a product. Sensing your moment, you arm yourself with a smile and utter the immortal line: 'Can I help you?' 'Yes', the customer replies, 'Can you explain to me the benefits of model K56G over L451 and whether the latest upgrade will still be available in five months' time?'

Congratulations and a bonus if you can adequately answer that question. The customer has probably been poring over Google for days, refining her product expertise to a level that only the most specialist customer service professional can hope to match.

This is one reason why some in the retail world have been lauding the benefits of bringing in virtual voice assistants into the retail space. Their access to such product knowledge could beat even the best-trained store assistant.

So far, so persuasive. Voice assistants are, after all, reaching critical mass in terms of customer awareness and sales – Amazon reported that its Echo Dot assistant was its best-selling product at Christmas, and voice assistant-enhanced products were all the rage at this year's major Consumer Electronics Show in Las Vegas.

But are we really saying that humans can be replaced by machines in this way, even machines with soft voices that have been programmed to dabble in a bit of humour? My feeling is that in the rush to get onto the latest digital bandwagon, some retailers may embrace this latest piece innovation too enthusiastically.

I believe there are three important reasons why we need to question this idea that voice assistants are the customer service Holy Grail.

Firstly, the voice assistant user experience isn't yet sufficiently refined to be able to cope with the kind of questions that most customers generally ask. The current generation of voice assistants work well as a kind of novelty for very simple requests like 'what will the weather be like tomorrow?' But the jury is out on whether they'd be able to handle a complex request like the one I began this article with.

Voice assistants are also limited in that they can only respond to a direct question. They can't (as a human can) notice a customer looking doubtful, annoyed or excited and respond accordingly. Neither can they offer product information unprompted – they can only regurgitate an online database of information in response to a very specific question. A good customer service person could get a sale by engaging with a lost-looking or decisive customer. A voice assistant would struggle to convert intuition into approachable, useful conversation.

This leads me onto my final reason why I think we are still far from the day when the machines replace staff in shops. The human touch in retail goes a long way – that's why friendly engaging retail staff are never out of work for long. Intuition and personal experiences enable human shop assistants to provide far more insightful recommendations than current iterations of voice tech are able to access. Being engaged by an illuminated screen or a disembodied voice will never achieve the same effect.

Maybe this is why the evidence is still contradictory on whether consumers will really like voice assistants beyond the first flash of novelty. A piece of research conducted last year found that 50% of consumers couldn't see the point of the technology. It may be relevant that this study was conducted in the UK – our friends across the pond, however, are embracing this kind of tech a bit more enthusiastically than us sceptical Brits, where, despite low usage for the moment, voice assistants are expected to reach 55% of US households by 2022.

The research did show though that young men under 34 and those earning over £41,000 are currently the two groups that are responding most positively to voice assistants. So it may be that retailers looking to appeal to these markets would show the greatest return on investment if they bring them into stores.

At this stage, I am neither advocating the wholesale use of voice assistants nor ignoring them entirely. Rather, the state of voice tech today demands a more considered approach that would use voice assistants selectively, in contexts that are appropriate for them and in conjunction with their human counterparts.

My belief is that a hybrid blend of the advantages of voice assistant technology with the unique qualities of human service will produce an optimum experience. This is about training and equipping human customer service teams with voice tech so that the humans take the lead in interacting with a customer while tapping into the voice assistants' ability to quickly summon data.

Perhaps the human assistants could use an Alexa-style voice assistants to find online reviews for customers, or offer them tailored vouchers. Finding the best blend of human and machine for each retail scenario  will be become the challenge but it will be worth the effort – better than taking a simplistic either/or approach.