AudioAudit

AudioAudit is an AI-powered auditing agent, enabling music copyright holders to scan their music catalogs & information for unrealized revenue potential by leveraging AI data analytical insights.

Comment

Inspiration

AudioAudit is an AI-powered auditing agent, enabling music copyright holders to easily scan their music catalogs & information for unrealized revenue potential through leveraging advanced AI data analytical insights. It will consist of a stand-alone platform and plugin to compliment existing music tracking services. There is an estimated $2.5 Billion dollars per year lost due to mismatched records, inaccurate data and incomplete information about copyright holders. There is an additional 10-30% of additional revenue growth potential estimated for businesses that can more accurately leverage data insights.

What it does

What sets our project apart from other music services that scan for copyright conflicts and missed revenue is that most systems only search for music in name, without cross-checking associated audio. This makes for incomplete data with regards to tracking royalties for copyright holders by not accounting for derivative works or covers. It also fails to provide valuable data on how songs are being used by the public and community in a way that adds potential marketing advantages by better identifying a song or artist’s target audience and/or trends.

Our project additionally considers a more comprehensive approach to data analysis by creating an evolving “Song Map”, where song IDs and data benefit from additional complimentary data added by other users in the community. We plan to use Zero Knowledge Proof to KYC users in order to verify their contributions to a music copyright (crediting, ownership etc) in addition to protecting sensitive contract and/or licensing data they might share. In this way we can help complete data and records for multiple stakeholders across a music composition or versions of a song without violating any privacy on behalf of those users.

How we built it

We built our project utilizing Typscript, Vercel, React and Javascript for speed with regards to designing the front end and to test out the input fields for the user experience as it relates to ingesting user song data, presenting the data and providing insights. AudioAudit would take the information provided by the user and utilize APIs to scrape corresponding information from various existing sources such as: DSPs (Spotify, Pandora etc), PROs and other royalty collection entities (ASCAP, BMI, MLC), Social media services (YouTube, TikTok) and more. We used FastAPI as a means to authenticate users. Given more time we would be establishing an AI model specifically trained on industry insights, user data and community user data.

Challenges we ran into

A major challenge that we faced was that many of the public APIs we wanted to use require API authorization and there wasn’t enough time to get our accounts authorized for this hackathon specifically. We also planned on using KindoAI to automate the processing of user data and also push insights back out to the user.

Accomplishments that we're proud of

We were able to distill a major data problem down to a simple solution for a wide variety of users. Our AudioAudit program suits the needs of copyright holders ranging from majors to indies and DYI creators. Typically only major corporations would have access to a service like this. We also found a way to coordinate the evolution of the quality of data insights provided across users through the concept of “Song Mapping,” while still managing to protect user privacy with our plan to integrate Zero Knowledge Proof processes.

What we learned

We did a lot of market research and learned that there are a lot of services scraping data, but not necessarily improving the quality ecosystem of data for the industry overall because they are keeping the results privatized and/or only using written data search methods as opposed to linking audio files comprehensively to account for derivatives and alternative works. We also learned that we have to apply to have access to certain APIs.

What's next for AudioAudit

For future growth, we would extend this model to account for a multitude of copyright royalty generating sectors including Film, TV, book publishing, games & more. Our business model would be to charge subscriptions for individual and enterprise accounts, with step-up services for advanced data analytical insights. We would also charge an administrative fee for any royalty revenue we assist in recovering or collections if they opt into those services. At 5% admin fee alone off of just estimated recovered funds of $2Billion missing annually, revenue annually would be 100 million. Expanding to other royalty generating industries (Book publishing, Photo/Video, Gaming etc) would provide access to an additional aggregate annual gross revenue of $444 Billion or $1-20 Billion in annual revenue for AudioAudit.

Built With

Share this project:

Updates