Founded by Naval Ravikant, cofounder of investment platform AngelList and all-purpose tech guru, Airchat is “Silicon Valley’s latest obsession”, according to an article in Wired earlier this month. But if you dialled the clock back to 2021, and replaced Airchat with Clubhouse, you could have written the same headline. While the audio-community app is now largely irrelevant, could Airchat be in danger of heading the same way?
“I still love the idea of the platform, but I have taken a step back from Airchat since my feed is not that buzzing anymore,” says Kesarwani from Bengaluru. “At first, talking to all my X mutuals via voice gave me a chance to build a better connection with them. However, a lot of my friends don’t seem too keen on the app because they are shy,” she adds. “And it’s no fun being there without friends.”
While it is too early to predict Airchat’s future, it prompts a wider question on the role of voice in navigating the internet.
VOICE ACTIVATED
Voice is the second oldest means of modern long distance communication, after the telegraph. However, in the digital realm, voice as a communication tool has not been able to capitalise on its promise despite early hype. Around 2018, companies began offering voice input for internet navigation—be it through search, command (via smart speakers like Alexa), or asynchronous communication by voice notes—in India. They saw the inclusion of voice-based features as an opportunity to onboard less tech-savvy people from non-metros, who could afford smartphones and access the internet through cheap data plans.Today, the voice note is a ubiquitous feature in consumer internet products— from social and messaging networks such as WhatsApp, Instagram and X to dating apps like Bumble and even the more professional LinkedIn. Some of them have audio playback speed controls to help you listen to voice notes at 1.5x speed or faster. Some web developers have built tech on top of WhatsApp voice notes to help people read voice messages. WhisperBot or Whisperize.me, an AI chatbot built by Maël H from Bali, claims to have helped people transcribe over 260,000 voice notes. “Devices with Alexa, like Echo smart speakers or Fire TV Stick, have been purchased by customers from across 99% of pin codes in India,” points out an Amazon India spokesperson in an email. One third of search queries on Google India happen via audio input, a company spokesperson told ET in June 2023. Ecommerce and hyperlocal platforms allow you to search for items on their app through voice. On the food delivery app Swiggy, you can also give directions to your address through a voice note.
SLOW SPEECH
Still, voice as a means to surf the net has yet to see wide-scale adoption. Over the last five years, Google search trends data in India for the phrases “Ok Google” and “Hey Google”, a proxy for voice searches, showed a peak of 100 in late 2022, but their popularity has declined to 17 and 60, respectively, now. In January, Google Assistant disabled a host of voice commands for actions like operating audiobooks and rescheduling events on Google Calendar as they are “underutilised”.
There was a 15.4% decline in global shipment of smart speakers in Q1 2023, year-on-year, as per IDC’s Worldwide Quarterly Smart Home Device Tracker report in June last year. According to a January 2023 article in investors.com, “the group inside Amazon responsible for the development of Alexa, called the Devices group, is reportedly running a deficit in the billions of dollars. And when Amazon announced plans to cut its workforce by 18,000, the Devices group responsible for the development and marketing of Alexa was hit hard”. Amazon India did not share a specific comment on this report while responding to an ET query.
Voice as a medium of input has not seen much traction on apps, says Prasanna Venkatesh, a director of design at a consumer internet company. “The numbers for voice search continue to be low in contrast to using text or auto-suggestions in search,” he adds. “I infer that voice is primarily used for search when the user is not able to remember the spelling or if the app supports localisation, like saying ‘nongu’ in Tamil for ice apple on Instamart.” Venkatesh is deaf and uses iPhone’s live-captioning feature to transcribe voice notes from friends who are not text-savvy. The voice recognition market is estimated to be worth $27 billion by 2026, according to a recent Statista report.
Meanwhile, there’s anecdotal evidence for the popularity of voice search and asynchronous communication among blue-collar workers; gig workers who are on the move and don’t have the time to text; and people who are unable to type in English and whose native languages are difficult to type on underdeveloped language keypads. It also serves as a vital tool for those with visual disabilities.
Then, why isn’t the voice commanding more attention?
A major limitation of voice as an input medium is “inaccuracy due to training sets,” says Soren Hamby, a product design and accessibility lead based in the US. “Most of these features do not provide accurate transcription based on gender, dialect and accent. One of my frustrations is that although I speak only English, I cannot be consistently understood by dictation processing because of my accent,” says Hamby, who has a visual disability and uses voice input “to dictate ideas, capture appointments, compose SMS or direct messages”. This recalls actorsinger Diljit Dosanjh’s clash with Amazon’s smart speaker Alexa over playing the songs of his choice in August 2020, indicating issues of comprehension of voice input, perhaps due to his Punjabi accent.
Also, “the use cases for voice search and command are often restricted to situations where you are looking for specific answers, like weather forecast or stock market update,” says Anand Jain, cofounder of CleverTap, a customer engagement platform. “I can ask my smart speaker to play a song on Spotify but not ask it to playback cells of an Excel sheet,” he says. Although voice integration is gaining traction across consumer internet platforms, its practicality remains confined. “We do anticipate a surge in the adoption of voice prompts over manual input on smart TVs,” says Jain. However, it is highly unlikely that you would say your password out loud while logging into a streaming service on the same device. Jain highlights how voice notes are often convenient for the sender and not the receiver who may have to replay the entire message a few times to catch missed details. The sensory overload and time-consuming aspect of voice notes have birthed an extremely vocal section that displays its aversion for voice notes on social media.
Dyutima Sharma, a trauma-focused clinical psychologist from Bhopal, says the anxiety and annoyance among recipients of voice notes can be attributed to non-verbal cues like tone and pace, and the receiver’s mental state. “For anxious individuals, a fast-paced voice note, sent without prior context, can heighten stress. While beneficial for senders, voice notes can strain receiver’s comprehension if these non-verbal factors are delivered poorly,” she says.
While the hype around voice may have subsided, there are some green shoots. “Many people use voice notes for privacy,” says Radhika Roy, a litigation counsel at the Internet Freedom Foundation. “Voice notes can be more privacypreserving than chatting interfaces, given the ease in dissemination and readability of screenshots. But some people avoid voice notes as there is a perception that they are more incriminating and identifiable than text messages,” she adds. In December 2023, WhatsApp announced disappearing voice notes to enhance user privacy. With threats around voice cloning and voice being used for biometric identification, Roy says users should be more cautious with their voice data and ensure that privacy settings of platforms do not authorise sharing of that data with third parties.
Meanwhile, Gaurav Singh, founder of Verloop, notes a growing trend among brands to enhance customer support in non-metros through voice notes on WhatsApp. Verloop is a customer support automation platform. “Currently, 5-8% of brand interactions on WhatsApp involve voice notes, but we anticipate it to grow,” he says. Singh highlights the ease of automating asynchronous chats compared with transcribing customer outreach calls.
Still, most consumer internet platforms may not prioritise advanced voice input features and just go with having basic voice search function, says Aamna Khan, a Bengaluru-based consumer retail entrepreneur. “Integrating voice-based features involves intricate technology, and determining their placement in the user journey isn’t a top priority for product managers, given the relatively low user adoption that is projected.” Khan, 31, relies on WhisperBot for transcribing WhatsApp voice notes, especially the five-six-minute narratives from her mother. She still checks Airchat daily, “reading along as she listens to multiple chats,” and hasn’t written it off just yet.
Voice navigation on the internet may not be mainstream yet, but at least it prevents you from having to “look between” two letters on your keyboard.
