Voice-based interfaces go mainstream: Are digital businesses ready?

By Surajit Agarwal Raj Koothrappali, arguably the world’s most loved Indian geek, fell for Siri way back in January 2012. The rest of the world has finally caught up with him. For us its Alexa, Cortana, Ok Google and of course, Siri. That’s because sometime this year than 61 percent of all smart phone users, spoke to the intelligent assistants residing in their phones. The increasing reliance on voice-based searches and commands is a direct result of the machine’s ability to accurately recognise words being spoken, understand their context and therefore serve up relevant results. In her recently released Internet Trends Report , Mary Meeker, Partner Kleiner Perkins Caufield & Byers, says Google is now able to recognize voice commands with 90 percent accuracy. Other voice engines range nearby. Badiu’s chief scientist Andrew Ng says “as speech recognition accuracy goes from 95 percent to 99 percent, all of us in the room will go from barely using it to using it all the time.” Ng also estimates that 50 percent of all web searches will be voice-powered by 2019.   Source: Kleiner Perkins Caufield & Byers Internet Trends Report Siddhartha Banerjee, Head of UI, UX Vodafone Shared Services, says “there is a wave of change coming and voice is just a part of it. We are moving towards the concept of zero UI. Visual clutter is going to be removed and the way we interact with computing devices is going to be much more intuitive and immersive.” Siddhartha sees the move to a voice-based interface happening on two fronts. “First there is the mobile phone, where we are already seeing a huge level of voice commands. Second I see this playing out in the home and the automobile. So the Internet of Things will be a big driver of voice-based interfaces.” The growth of voice of course, comes with a number of caveats. Recognition human speech is getting better, understanding context not necessarily so much. Natural Language processing, deep machine learning, neural networks and its ilk are wonderful PR copy, but still have a way to go in terms of real-time, real world implementation. Ripul Kumar, Director User Experience at [24]7 says hybrid voice interfaces “are the stepping stone to full voice-based navigation systems. We already have products that use a mix of voice and web pages to help reduce customer service costs. The next step is fully-voice based interactive systems in a limited environment. For example, any location-based taxi service is a perfect fit for a voice-based system. The system only has to learn a certain number of commands sets. It only needs to parameters like the ‘to and from’ information, location and the time. This is easy to manage. It gets more complicated if a big e-commerce platform selling thousands of products tries to implement a voice based interface.” Challenges in designing a voice-based interface Language: There are only a limited number of languages recognized by voice-based search assistants. In India it is only Hindi. Others will get there, but in the meantime we are talking about a fragment of the total smart phone population talking to their phones. Short-term human memory: Yes, it is human memory that is a problem. If you want to understand why think back on the last time you used a bank’s Interactive Voice Response (IVR) system. How many of those eight or nine options did you remember? Probably one or two. Our short-term memories are pathetic. On top of that says Ripul “Voice-based systems are linear. That means for a query we could be inundated with responses one after the other. Think of the number of responses that appear on a Google search and imagine that coming to you via a voice output. Temporary disability: Voice commands are frequently used when the user is unable to use any other type of input. So while driving a car, or in standing in the metro and having only one had free. In these situations ambient noise plays a big part. This could compromise the software’s ability to understand and recognise the command. This is also the time users would not like to presented with multiple options. So if the query is for a “Chinese restaurant nearby,” UI developers have to design it in a way that despite the question the answer should be limited to one or maximum two options. Ripul says the key is “for UI designers to take on the onus of providing the correct answer, rather than throw a number of almost relevant answers at the user. Designers have to be cognisant of the fact that there are certain things that cannot be done in an audio interface. Where a human being expects a single or two discrete outputs that is where it will work, more than that it will become a burden on human memory. Small and discrete tasks with no relation to subsequent tasks, those are fine. Marketing for voice-based interfaces Sneh Sharma, Founder and CEO of Digital Agency Ittisa, says the way we process content for websites and mobile sites will need to change. She lists the following elements to ensure voice-based searches are able to find your product. Optimising for SEO: Brands will have to start optimising for long-tail keywords. Schema markups which is an addition to HTML web coding can help search engines extract answers from your website. It provides additional information, which allows the search engine to understand the context of your search better. Make location specific data readable: 40 percent adults and 38 percent teens use voice search to get directions, which is why it is important for brands to make location-specific data readable. If your company has locations open to the public, it’s essential that these locations be found via Siri, Google Voice Search and other voice-activated search technologies, which beyond your site’s spot information, relies on unified citations. Therefore, it is imperative that there are no discrepancies between your addresses online. Use FAQs to show up on searches: Voice search is built on the premise that a user asks a question to which there is an answer. This makes FAQs a very important part of your website and therefore content that is framed around the 4Ws and 1H (who, what, why, where, how) is essential. Conversational phrases: The important thing to keep in mind while building content is that is has to be framed in a conversational manner. High quality, well-researched content with credible authorship will show up on search engines more often. Provide a seamless experience: Brands will have to start using voice search to receive direct orders from customers. For example, Dominos launched their digital assistant Dom through which customers can directly place customised orders. Finally how much time do you have to get your digital properties voice-ready? Siddharth reckons voice-based search will be mainstream in India within the next year. So that means you have to start now.   The author blogs about digital marketing.