Voice AI technologies have come a long way since its first primitive iteration back in 1961.
At the time of its release, IBM Shoebox could recognise just 16 words and 9 digits but laid the early foundations for what would become a multi-billion-dollar industry.
A number of developments built on this idea, with probably the most memorable being Microsoft’s Clippy, introduced in 1996. No matter your opinions on this quirky, yet slightly intrusive little animated paperclip, these technologies were starting to show us how natural language in text could be decoded and responded to by a series of set responses, programmed to recognise patterns in text in order to solve user issues.
Since Clippy, developers have stuck to the principle that Voice AI technologies should only introduce themselves when called upon, rather than interrupting and questioning if you need their assistance at every given opportunity.
This idea has followed ever since with the creation of what we now recognise as modern Voice AI with the likes of Siri, Alexa (don’t read this out loud), Google’s Cortana and so on.
Sales and the prevalence of such products have only gone from strength to strength. In the U.K alone in 2017, commercial sales made by consumers utilising Voice products amounted to £200mil in sales with that estimated to rise to a whopping £5bil by 2022*.
The industry seems to be on a similar trajectory to that of the smartphone. Very few people owned a phone that was constructed predominantly of a single screen before 2007. But 14 years on, you’d be hard pushed to find someone using anything else; if you’re working from home still, I’m sure you don’t need to look too far for your own personal Voice AI device.
Voice is the easiest and most natural form of communication we have in our arsenal. Combining the ease of ordering online with the simplicity of voice communication, while not having to even touch your device, makes for a marriage made in consumer heaven.
But what of the use of Voice AI by enterprises and in the world of business communications?
Firstly we should address the rather large elephant in the room: Covid.
At the time of writing, there’s justifiable hope that we’re beginning to turn the page on a pandemic that has put many things on hold the world over.
But the aftershock of such an event will surely remain for a long while after its resolve. Simple actions like touching surfaces and an increasingly concerted focus on hygiene will remain ever-present. Lots of the things we use in the office would have been used by at least one other person, at least one other time; with the increased adoption of things like hot-desking and open-plan office design, never has this been more at the front of people’s minds.
Along with Voice AI’s simplicity, and as long as it remains reliable, it could offer employees an alternative that addresses understandable, lingering concerns around bacteria transmission.
“Voice is becoming the universal remote to reality”James Vlahos, Talk To Me.
These concerns may subside as time passes, but for the initial period, where workers begin to transition back to some form of office working (whether full-time or part-time), an opportunity is opening up for Voice to become a mainstay in our working lives.
Businesses are often swamped with basic queries with Sales and Customer Service teams having to take considerable time to offer the same answers time and again, costing a lot of time and manpower.
Voice bots have become increasingly utilised by companies on websites and for automatic email generation to deal with basic customer queries, as well as ticketing systems for internal communications.
Gartner.com’s VP Analyst, Van Baker predicts that, by 2022, 70% of white-collar workers will interact with conversational platforms on a daily basis.*
This helps relieve the workload from teams that deal with the response process so they can focus on more complex queries. But this is far from being the norm for many businesses.
Voice automation doesn’t just offer the ability to free up time but can also provide a faster service as well as open up significant scalability opportunities.
Many video meeting providers also offer the ability for users to analyse speech patterns and highlight key moments from meetings.
Voice AI is gradually becoming a key component of the video meeting offering. This offers participants key insights to make sure their meetings are focused and as little time as possible is wasted on topics unrelated to the meeting.
These sorts of meeting analytics also provide a handy tool to trainers and management to ensure frontline staff are communicating the appropriate message. Whether it’s an important sales pitch, customer advice call or communication of brand values, these things can be easily understood and improved through Voice AI analytics and speech pattern recognition.
Some Potential Pitfalls
But for all the upsides, Voice AI can open up businesses to potential risks.
Voice AI needs to continue to develop for the benefit of its users; AI needs to remain practical, and not just a side gimmick, in order to encourage adoption in a more formal business environment.
It’s all a bit of fun when you ask Alexa to play Smells Like Teen Spirit by Nirvana but instead ends up adding a bottle of Gordon’s to your shopping list. But having a support chat bot never reach a definitive answer for a customer, or an important meeting fail to connect because it doesn’t recognise the host’s voice, brings with it serious headaches and potential embarrassment for the company concerned.
This addresses one of the key flaws in computerised speech recognition. Getting an AI bot to recognise patterns in speech is difficult enough, but throw in context and semantic fluidity, and you have a recipe for misinterpretation.
Let’s take this one example using the verb ‘Run’ which has over 200 potential uses.
A boat can run aground; you can run down the street; you can run for mayor or run a business; cars can run on petrol or I could ask you to keep the engine running while I run into the shops. Of course, AI is getting better at determining these differences, but really understanding each nuanced use is still a human art. This is before we add in idioms and metaphors into the mix.
Consistency and reliability are key for Voice AI to be considered useful within a professional setting. Mistakes and teething problems are normal in a technology’s infancy, but if this becomes more than a one-off, companies could reject the technology entirely despite its upsides.
With the growth of a new technology, comes a greater concern for security.
Whether that’s confidential meeting recordings being left available to anyone within the company, or external intrusions infiltrating meeting systems, the security of Voice AI technologies must match the benefits to ensure its long-term integrity.
Much like the point made above, it takes years for technologies to build a solid positive reputation, but only one serious incident for people and businesses to dismiss it entirely.
Voice is the easiest and most natural form of communication. Never has this been easier to utilise for business purposes through the numerous video meeting providers available today.
Through Voice, personified AI is gradually staking a claim for becoming the next big technological phenomenon of the 21st Century.
Through Voice AI, humans are slowly crafting additional personal assistants, diary managers, meeting analysts, customer service helpers, and many others, in order to free up time for other more complex and/or time-consuming tasks.
There’s still a long way to go from the image we may have in our heads from 70s and 80s sci-fi films, but fiction is slowly, but surely, becoming reality.
This technology is at a comparatively early stage within the consumer space – compared to the likes of the smartphone – but it won’t take long before this bleeds through to find its own business use and becomes an increasingly prevalent presence in our working lives.
Talk To Me – James Vlahos
Gartner.com | Trends
The History And Growth of Voice AI and Search | Richard Liu
Why Voice AI is critical for enterprises in a post Covid world | Arun Santhebennur