When it comes to new publishing opportunities, voice interfaces are right up there alongside AI, augmented reality and blockchain. But voice interfaces come with their own set of unique challenges and more than just a few concerns. To help navigate this new frontier, WNIP has put together the ultimate guide to voice interfaces for publishers.
Voice Interfaces – the Complete Guide
What exactly are voice interfaces?
Technically speaking, they should be referred to as voice-user interfaces. That’s VUI to you and me. VUI makes human interaction with computers possible through a voice platform to initiate an automated service or process. Think Tony Stark’s JARVIS, but without the sarcasm.
Amazon uses Alexa, which was inspired by the VUI on board Star Trek’s Enterprise. If you have any doubt as to Amazon’s lofty opinion of their prospects, then the 5,000 employees working on Alexa should set you straight.
Apple uses Siri, which has come under fire for its poor AI despite Apple’s early lead in the market. In fact, the entire project has been rife with divisions and political infighting by all accounts. No wonder then that Apple recently poached Google’s AI chief John Giannandrea to run its machine learning and AI operations.
Google, meanwhile, employs the creatively titled Google Assistant to interact with their products and services.
The industry view: Google Home just about edges Amazon’s Echo, namely because of its strength in search. This is somewhat predictable, considering Google has been in the business of cataloging information since Jeff Bezos was doing little more than shifting copies of Harry Potter.
What is the global penetration?
Huge. And growing. According to market research group Forrester, the installed base of smart home devices in the US alone is set to reach 244 million in 2022. They predict that smart speakers, including Amazon Echo, will account for 68% of the total installed base of smart home devices that very same year. Forrester expects this to accelerate on the release of the next generation of smart speakers, which will be combined with emerging smart home systems.
Sales have been dominated by the US and UK markets, with household penetration standing at 6% and 3% respectively. A recent Gallup poll found that 22 percent of Americans already use devices like Google Home or Amazon Echo.
What publishers need to know
There is a land grab to own skills
In the landscape of VUI, skills are the same as owning a web domain. It’s essentially an instruction to a voice assistant about a specific topic.
When it comes to voice-assistant capabilities, skills can be split into two categories. The first is branded skills which – perhaps predictably – are linked to your brand, and could not be owned by any other company. Skills such as TED’s ‘play the latest TED Talk’ action and the Wall Street Journal’s ‘What’s News?’ fall into this category.
The second category encompasses more generic skills. These could be things like “Alexa, give me the latest publishing headlines” or “Okay Google, give me the latest finance news”. Ownership of these generic skills gives you sole ownership over entire categories, creating a first-mover advantage in the market as brands race to capture skills before they are gone. This can make things difficult for brands looking to harness a particular generic skill. However, there are opportunities for publishers to harness market-specific skills in both the generic and branded categories – like most things, it’s just about finding the right target.
Optimising for Voice Search could boost revenue
According to a recent study conducted by Consumer Intelligence Research (CIRP), Amazon Echo customers spend 66% more than average Amazon customers. Amazon Echo customers spend, on average, $1700 per year, while their counterparts spend $1000. Members of Amazon Prime spend an average of $1300 per year. This means that Amazon can now afford to sell Echo devices at a lower price than originally planned, occasionally taking a loss on devices to gain a greater share of consumer spending. For publishers, this implies that optimising for voice search could result in a revenue boost.
How have publishers been using VUI? (The Good)
Reinventing Customer Experience
In his keynote address at the CMO Digital Insight Summit, Amazon’s general manager of Alexa Skills Fabrice Rousseau spoke about reinventing customer experience through voice technology. “When we moved from desktop to mobile we didn’t bring the desktop experience to mobile, we invented a very specific mobile experience,” he pointed out. “When you move from mobile to voice don’t bring your mobile experience. Just invent an experience that is unique to voice.”
Building your brand with skills
It’s important to note that, as far as most publishers are concerned, VUIs are still in their experimental phase. Despite early successes with branded skills and flash briefings, VUIs still operate at a fairly low level – following commands to play music or read news headlines, for example. That said, many publishers are already working on plans for expansion, and with the land-grab to own skills still underway the future successes of VUI belong to those who move first.
Most major news outlets have at the very least developed a daily briefing, with varying degrees of sophistication and success. NPR’s News Now is widely considered the leader of the pack, with five minutes of news delivered by an NPR broadcaster. One factor that gives News Now the market edge is that it updates every hour, while flash briefings of a similar quality from the likes of BBC News and CNN aren’t quite so up-to-date. In addition to flash briefings, outlets such as The Guardian have made all of their news, reviews, podcasts, and comment pieces freely available on voice platforms, while the Daily Mail has instead developed a branded ‘news on demand’ service that can only be accessed by digital MailPlus customers paying the £9.99 monthly subscription fee.
Aside from daily briefings and newscasts, some publishers are developing more brand-specific skills and content. HarperCollins Christian Publishing, for example, has developed a ‘Devotionals’ skill featuring short uplifting passages read by authors as an extension of their existing social media ‘daily devotionals’. Other publishers have combined interactive elements with their regular VUI programming, with both The Washington Post and Financial Times launching quiz skills based on news coverage and public figures.
In 2017 BBC R&D began to experiment with native VUI content, and released an ‘interactive science fiction comedy story’ called ‘The Inspection Chamber’ for Amazon Alexa and Google Home. In 2018, they plan go further down this route, and will look at developing more children’s VUI content as part of their remit as a public service. For other publishers, such as Bayerischer Rundfunk, the next logical target is to build a journalistic dialogue between users and voice platforms. But what would this ‘dialogue’ look like? For one, it would allow customers to interrupt a newscast to learn more about particular items and dictate the direction of their own newsfeed.
The personal touch
One issue that publishers have been running up against with VUIs is voice-assistants’ total lack of personality. At the Financial Times, head of FTLabs Chris Gathercole and his team have been using Amazon Polly to convert existing text articles to audio that is then delivered by ‘Artificial Amy’. While ‘Amy’ learns quickly and is cost effective, her lack of human inflection can be off-putting and can steamroll the humour or nuance of a piece.
On top of the comprehension issues this obviously brings up, automated voices are often either boring or straight up disturbing, both of which put users off. Gathercole believes that a blend of artificial and human voices could temper the issue, with a voice actor reading parts of the text and a computerised voice contributing further snippets.
Another solution to the problem, as Google’s Peter Hodgson pointed out at the Smart Voice Summit in Paris earlier this year, is to build a brand-appropriate persona for your voice assistant. Quartz, for example, originally used Alexa to deliver their Daily Brief but soon found that they had more success using a variety of voices in the more conversational tone they had established with their mobile app. Quartz’ Alexa newscast is now read by a pair of AIs called Kendra and Brian, which present headlines in a more playful tone than Alexa is capable of. Similarly, the Telegraph initially used Alexa’s pre-programmed voice to deliver their ‘5 by 5’ Google Home show, but now have their journalists read and discuss the news. Since they made the switch in late 2017, the Telegraph has seen an increase in the number of subscribers tuning in.
Where to begin? There are all manner of privacy concerns surrounding the ownership of devices that are essentially constantly eavesdropping on your home. While VUIs can bring to mind the sleek aspirations of sci-fi movies and futuristic projections, there is also the not entirely unfounded fear that they could very easily be used to monitor and manipulate ‘Black Mirror’ style.
Customer advocacy group Consumer Watchdog shed light on some of these concerns in a recent study of new patent applications for Amazon Echo and Google Home. “These patents show that smart devices target moments in between screen time to monitor sleep habits, listen in on dinner conversations, and track when users shower. Access to this data can flesh out Google and Amazon’s profiles of their users in order to help them more accurately server targeted ads,” the study claims.
Consumer Watchdog’s assessment is unequivocal; “Digital assistants like the Amazon Echo and the Google Home greatly expand the collection of personal data, magnifying the risk that someone will learn something about you that you would rather keep private,” they conclude. So, there isn’t a whole lot of wiggle room for interpretation here. In an article written by Sapna Maheshwari for the New York Times, Consumer Watchdog president Jamie Court doesn’t skirt the issue; “When you read parts of the (Alexa) applications, it’s really clear that this is spyware and a surveillance system meant to serve you up to advertisers,” he says.
The thing is though, none of this is illegal. While future developments in VUI seem to have enormous potential for corruption and misuse of data, it is still early days. Alexa was launched just over three years ago in November 2014, and we have not yet seen the full scope of the technology’s potential. As VUIs – and our understanding of their implications – grow and change, there may be safeguards put in place to prevent misuse of data. With the recent issues surrounding companies such as Facebook and Cambridge Analytica, however, these concerns cannot be dismissed out of hand.
The Final Word
While VUI is still in its infancy, many publishers are making rapid advancements and future successes will belong to those who move quickest. Despite user privacy concerns and some slightly creepy hardware glitches, VUIs present an excellent opportunity for publishers to reach a wider audience in new ways and have the potential to increase revenue.
June 12th: Get Alexa on Your Apple Watch With This App
At a glance: There are times where you might not be around your Amazon Echo to access its features, especially smart home features. Voice in a Can works on its own without your phone, so you could be out for a run and lock your doors or start your coffee machine. Basically, it gets you one step closer to that Inspector Gadget lifestyle we’re all working toward.
At a glance: Given the challenges that many people have had with the accuracy of Siri’s recognition, this more simplistic approach is a good fit for Apple. Essentially, you’ll be able to do a lot of cool “smart” things with a much smaller vocabulary, which improves the likelihood of positive outcomes.
At a glance: Amazon has launched a beta version of an interface called CanFulfillIntentRequest, which will enable Alexa to suggest the perfect skill for users when they don’t know which one to ask for.
At a glance: A team of academic researchers has tested smart-home assistants Amazon Alexa and Google Home, finding it possible to closely mimic legitimate voice commands in order to open a rogue app instead of the legitimate one, thus hijacking the connection. In short, voice-squatting.
At a glance: A Portland family contacted Amazon to investigate after they say a private conversation in their home was recorded by Amazon’s Alexa and that the recorded audio was sent to the phone of a random person in Seattle, who was in the family’s contact list. In short, Alexa is still grappling with fundamental issues of privacy.
At a glance: There are some important considerations developers should take on board to overcome the unique challenges posed by delivering a positive brand experience in a voice environment. To add genuine value and avoid creating vanity apps, it’s imperative that businesses put the needs of their customers first.
At a glance: Many commentators have suggested that “Duplex” is not only strange but entirely unethical, and that it could signal an important moment in the acceptance and use of artificial intelligence.
At a glance: Virgin Trains has become the first train operator in the world to sell its tickets through Amazon Alexa. Rail travellers can now book Advance Single tickets and sort their trips with one simple voice-based transaction with payment completed through Amazon Pay. Rumours that the VUI suffers from delays are unfounded.
At a glance: ‘Sound-to-Meaning’ startup SoundHound has received $100 million in new funding from Chinese tech giant Tencent (and others) to give every brand a voice. And not just any voice: their own voice.
April 27 2018 Study finds Google Assistant is smarter than Alexa
At a glance: For the second year in a row a study has found that Google Assistant is ‘smarter’ than its competitors, generating the most correct responses to questions across the board. Somewhat surprisingly, second place goes not to Alexa, but to Microsoft’s Cortana.
April 26 2018 Amazon announce new skills for Alexa.
At a glance: Amazon are developing new capabilities for Alexa, including an internal memory that will allow Alexa to recall previous conversations or information you have provided. They will also improve Alexa’s context capability, so that follow up questions don’t have to be asked separately.
April 25 2018 Amazon launches Echo Dot for Kids
At a glance: Amazon have launched a new Echo Dot product specifically aimed at children, which comes equipped with new Amazon FreeTime Unlimited software. Amazon FreeTime Unlimited features parental controls, a bedtime limit, and child appropriate programming and responses. It also responds positively when requests include the word ‘please’.
April 24 2018 Amazon make their Alexa app iPhone X compatible
At a glance: Amazon have updated their Alexa iOS app to be optimised for iPhone X, including bug fixes and ‘performance enhancements’.
April 23 2018 Google update their podcast capability
At a glance: Google have updated their podcast capability for VUI devices, so that users can begin listening to a podcast on your phone and then pick up where they left off later with Google Home.
At a glance: LG have joined the ranks of the electronics companies whose devices can support both Alexa and Google Assistant, preventing customers from being locked in to one voice assistant.
16 April 2018 Adobe acquires voice interface platform Sayspring
At a glance: Adobe has announced that they have acquired Sayspring, a startup that helps clients develop voice interfaces for Amazon Alexa and Google Home.