Monetizing audio is the new competitive advantage

AI transcription is turning audio from a costly manual bottleneck into a scalable competitive asset. Here's why forward-thinking organizations are taking notice.

Audio, and the emerging trend of audio as a data set, is a valuable but underutilized resource. Until recently, the transcription of audio was expensive and often slow due to manual processes. In short, it was only as fast as qualified humans could properly transcribe it.

That’s changing thanks to major breakthroughs in AI transcription over the past few years that are enabling decent-quality transcription that can be completed by a machine almost instantly for a very low cost. While human-edited transcription is still the best way to ensure a near-perfect end result, AI transcription has improved to around 80–90% accuracy on most audio.

In fact, the global voice recognition market size was forecasted to grow from 10.7 billion U.S. dollars in 2020 to 27.16 billion U.S. dollars by 2026.

Transcription’s role in monetizing audio

Audio in its native format is unwieldy and hard to search. Have you ever tried to go back through an audio recording to find a specific spot? It’s painful. You just can’t do that much with audio until it’s been transcribed. There are several main types of transcription:

• AI transcription, which is fully machine generated

• Human-edited transcription, also called human-in-the-loop or HITL, which starts with an AI transcript and uses humans for the “last mile” to correct errors

• Fully human transcription

Once audio data has been transcribed, it becomes searchable, multi-faceted, and actionable. Machine learning can be automatically applied to the transcript to enrich, link or categorize the content. The text can be scanned by search engines to improve SEO on your website. Transcription can be integrated into workflows to improve efficiency, adhere to compliance guidelines, or reduce administrative burden on staff. No matter how it looks in your organization, quality transcription is the first step in monetizing audio data.

Kensho offers two unique services for transcription: AI and human review (HITL). What resonates most with our customers is Kensho’s versatility within the business and finance audio space. Our in-house teams ensure flawless accuracy, while also guaranteeing consistent delivery in specified times (6 hours — 72 hours). On top of this, we offer custom formatting and style guides for our users who would like to have specific branding on their transcripts. Contact us to talk about Kensho Scribe’s AI or human-edited transcription.

Use cases for monetizing audio

Many businesses that record or collect audio as part of their workflow could be a good use case for monetizing audio. Here are a couple of examples:

Expert network firms

Expert networks, which arrange calls between investors and experts, need high-quality transcription because their calls serve as critical primary research that is not readily available on the market. There are actually two ways within the expert network business model that focus on audio monetization: transcripts are sold directly to a client’s customer base or sold via a subscription to an online portal browsing service.

Expert networks tend to have the highest volume of audio consumption within the professional services industry. For the largest clients, their yearly hour ingestion can range from 3,000–10,000.

These firms need a service that can handle massive volumes of audio data, produce highly accurate transcripts at specific delivery times and also have customizable/formatting features that fit their own branding.

Podcasting

Podcasting is an area Kensho is familiar with both internally and externally. We transcribe several of S&P Global’s podcasts including NextInTech and The Essentials Podcast. Externally, we transcribe audio for a finance-based podcast with over 200,000 monthly downloads.

For podcast platforms or series, having transcripts available on their website provides several benefits: new inbound traffic increases by an average of 5% and new inbound traffic via search increases up to 10%, both directly attributable to new audio transcription pages. Additionally, podcast transcripts are fantastic for marketing content and unlock several avenues for brands to capitalize on, whether it’s social media, newsletters, videos, blog posts, etc.

A case study:

One example of our technology is highlighted through our Tegus case study. To ensure that Tegus got an unbiased look at how different vendors compared in terms of quality, the team decided to do a transcription “bake-off” by sending Kensho Scribe and other top transcription vendors real call recordings and comparing the outputs they received back.

The returned transcripts were evaluated for sentence-by-sentence accuracy as well as the number of inaudibles, words that the service couldn’t transcribe at all. During the bake-off, they also tested each vendor’s ability to handle Tegus’ massive volume needs by moving all of their transcription work to one vendor for a period of time and then onto the next for the same length to compare how each did under the increased volume.

Polly Benassi, Director of Compliance Operations at Tegus, said, “We tested quite a few and looked at them side by side. And Kensho was the clear winner.” On one transcript example, Kensho returned with only one inaudible where other competitors had many, including one vendor which returned the same transcript with 52 inaudibles. Polly described how some of the vendors would mis-transcribe one word or one sentence and that would change the entire meaning. “I wrote up a side-by-side comparison of a paragraph and just how many sentences were accurate compared to the audio. Kensho’s were 95% accurate.”

When it comes to monetization of audio, the latest technology advances mean there’s a world of opportunities for businesses that have audio. Call centers, for example, could do so much more with their audio if they transcribed everything and used natural language processing (NLP) to spot issues and trends. These advancements in AI mean that anyone with audio can now quickly and cost-effectively transcribe as part of their normal business workflows and use AI tools to automatically and systematically enrich and analyze the content.

What type of audio does your business produce? Talk to Kensho about how our tools can help you monetize your audio.