“Voice Search” is a term used for both voice-enabled search to retrieve information and find phrases in audio files. Searching for information with voice inputs is also known as “search by voice”, “voice-activated search” or “search with voice”. It allows the user to use a voice command to retrieve information from the Internet, a website, or an app. Search powered with voice also enables open-domain keyword query, and that query is provided to the machine by talking rather than typing on a hard or soft keyboard.

What’s Search by Voice?

Graphic UI for Google search by voice, also known as Google voice search, voice-activated search or search with voice

Google Search Engine introduced the Search by Voice feature for the web a decade ago, but since then it’s only available for the Google Chrome browser. To use Google’s search by voice, a user clicks on the microphone icon, the push-to-talk button on the right-hand side of the search bar. Push-to-talk button activates the voice-enabled search engine and it starts listening. Users dictate the keyword or phrase, instead of typing. It’s quite similar to the interaction with Google Assistant. Instead of saying “Hey Google”, users click on the push-to-talk button to initiate the search. The rest of the queries such as “what’s the weather?” or “what does automatic speech recognition mean” are processed similarly.

Search powered with Voice AI in Numbers

0%
of the online population using search by voice on mobile
(Google)
0%
of Americans used voice daily to search rather than typing
(PwC)
0%
of Americans prefer voice to search over type to search
(PwC)

The user adoption of voice, considering the benefits such as speed, ability to multitask, and convenience, especially on small screens such as mobile phones, is not surprising. Users can speak 3 times faster than they type, and can find products such as “Red Women Sneakers Size 10” much easier than typing or even worse filtering on several drop-down menus or checkboxes.

Challenges of Google search with voice

Despite the convenience of voice for mobile and web applications, voice-activated search has gained popularity with voice assistants such as Alexa, Hey Siri and OK Google when they’re integrated into smart speakers and mobile phones. It’s mainly because these voice assistants have been the only option that offers a truly hands-free voice experience. Even Google search enabled with voice AI for Chrome doesn’t support a wake word. Users need to click on the push-to-talk button.

When it comes to limitations, Google is the least bad. Some solutions offer hold-to-talk buttons. Hold-to-talk buttons require users to hold the button throughout the search query or input, making it even less convenient, especially when you think of a salesperson adding notes to CRM while going to another meeting.

How to build a truly hands-free voice-activated search experience

Picovoice’s Porcupine Wake Word is the only engine that works in web and mobile web browsers. To back this claim and give developers a head start, the Picovoice team built a demo Voice AI Browser Extension. Even Google Search on the web can be a truly hands-free experience.

Platform challenges to add voice to search experiences

The chart shows Speech Recognition API’s platform support for web and mobile web browsers

Limited platform support is another limitation that hinders the adoption of voice search on the web and mobile applications. It makes developers’ lives difficult and requires enterprises to allocate more resources to voice projects. A user may access a web or mobile application through Chrome, Safari, Firefox, Android or iOS, and expects the same experience. However, adding voice to these different platforms requires the use of multiple SDKs or APIs, hence more developer time.