Top Six Use Cases for Automatic Speech Recognition (ASR) in Social Media

Social media platforms have truly changed the landscape of how people all over the world engage with each other, keep up on news, explore their interests, and more.

And their reach and impact continues to grow. Every second, approximately 6,000 tweets are sent and, as of this writing, 97,306 YouTube videos were being viewed in one second. With this much content being created, it’s hard for social media platforms to keep up. This content has the potential to be a treasure trove of information, but with more and more uploads taking the form of video and audio content, much of this content has been somewhat opaque to the platforms that host them.

That’s where automatic speech recognition (ASR) and Natural Language Understanding (NLU) comes in. By taking audio and video files and transcribing the speech that occurs in them, social media companies can gain access to a whole new source of information about what’s benign discussed on their platforms and provide additional value to their users. And, the more accurate the transcriptions and classifications, the more effectively social media platforms can take action based on this information.

In this blog post, we’ll look at some of the top ways that social media companies are using transcribed audio in their businesses today, as well as the ways that ASR and NLU solutions can drive increased revenue for social media platforms.

Social media companies use automatic speech recognition and speech-to-text tools for a variety of purposes. Let’s take a look at six of the most common ways that this technology is used.

Speech-to-text solutions allow social media platforms to add captioning to audio and video that’s uploaded by users. These captions not only support people in the deaf and hard-of-hearing community, but also anyone who is browsing social media in a place where they can’t listen to audio. Captioning like this serves to increase engagement for posts, even if people can’t hear what’s being said.

Social media companies can use automatic speech recognition and NLU features such as Language detection, Topic detection and Entity Detection to provide monitoring services to customers about their products and services. This information can be used to improve the user experience and make changes to the products and services based on user feedback. It can also be used to moderate the content that’s being produced and uploaded to ensure that it confirms with site guidelines and any relevant laws and regulations.

Social media companies can use speech-to-text tools to help with ad targeting. In the past, anything that was uploaded in the form of a video or audio was largely opaque to the platforms. While a social media platform could target ads based on text posts, this wasn't possible for video and audio beyond the title and any metadata included as part of the upload. But with ASR solutions, social media services such as podcast platforms can target ads based on the content of audio and video posts themselves by using features such as Keywords, helping to better monetize the platforms.

Social media companies can use automatic speech recognition to improve the search functionality of their platforms. In the past, as with ad targeting (discussed above), search had to rely on titles, descriptions, and metadata to help users find the content that they're looking for. But with speech-to-text functionality, videos can be transcribed and served up for search, making it easier for users to find the content they're looking for, regardless of its format. Developers building social media applications with search functionality based on audio content can now leverage ASR features that identify terms or phrases by matching acoustic patterns in audio.

Better understanding their users is widespread use of automatic speech recognition and speech-to-text tools. By analyzing the audio and video content that users upload, social media companies can get reliable information that can be used to surface trends and insights to the business, and/or automate downstream business process workflows. One example is to route posts for content moderation to the appropriate team based on language detected or filtering profanity in real-time.

As mentioned above, the main benefit that ASR tools provide to social media companies is the ability to unlock what’s happening in audio and video that’s been published on their platform.

• Save Money — Many of the use cases above support saving money. Whether it’s optimizing customer service calls or better understanding customers, deploying speech solutions can save social media platforms big bucks. By providing Closed Captioning services to their customers you can also enable your customers to demonstrate compliance and avoid regulatory fines.
• Generate New Revenue — By providing assistance with ad targeting and highly accurate search, ASR solutions can increase the revenue generated by social media platforms.
• Increase Productivity — With customer insights, ad targeting, and improved search ASR solutions can increase productivity for social media companies. Instead of having to do these things manually, social media platforms can rely on software solutions to enable those tasks. For example, Deepgram Customer This allows them to focus on revenue-generating areas such as additional ad packages, new analytics offerings, new products and features, and more.

As we’ve seen, ASR solutions can help social media companies in many different ways, such as improving customer service, saving money, and better monetizing their platforms. In the future, we can expect social media companies to use these technologies in even more ways to improve the user experience.

If you’d like to see how Deepgram can support your speech-to-text needs, you can sign up for a free trial of Console and get $150 in free credits to try us out. Have questions? Reach out to our sales team; we’d be happy to talk you through your use case and see how we can help.

If you have any feedback about this post, or anything else around Deepgram, we'd love to hear from you. Please let us know in our GitHub discussions .