AWS Startups Blog

Aiconix Transforms the Way Companies Use Their Content

Analysts estimate that by 2021, videos will account for over 80% of consumer internet traffic. To stand out companies and brands will need to maximize the value of their content and understand how a story, a posting, a video can be optimized to leverage consumers interest. Identifying and owning the relevant data to achieve this is just the first step in the new workflow. Aiconix, named one of the 30 most promising startups in the German speaking regions by Forbes magazine, is doing just that.

Aiconix’s platform aingine extracts data for customer’s from their audio-visual content and correlates these insights with data from analytics to finds patterns, predict user behavior, and give recommendations on how to change the dramaturgy of a video to reach the desired target group and reduce the bounce rates.

Imagine you are writing an article and in order to make it multimedia-based, and you have to manually search your media archive for images and audio-visual files relevant to the topic. This is very resource- and time-consuming. The aingine from aiconix offers the ideal solution for this: with the aidvice feature, texts are semantically analysed and suitable media files from your archive will be automatically recommended for use. Depending on the application, aidvice also automatically establishes thematically related links across all media types and makes corresponding recommendations, such as topic related podcasts or graphics.

Another video production use case is if you are looking for already produced material on a specific person in the image and video archive. This person, however, has only recently become the subject of increased public interest and has therefore not yet been named in your archive. A manual search for this person in your archive would cost a lot of time; however, with the features faice of aingine, this is possible without much effort. Faice automatically tags each person to be seen in pictures and videos with an independent ID. You only need four images of person X to find all other materials by ID. The ID is then simply replaced with the name of the specific person.

Highly appreciated among Slack users is the Speech-to-Text feature of aingine from aiconix which automatically transcribes a video interview or discussion. The slackbot acts as an interface between the user and aingine. You have the possibility to drag and drop audio and video files in several languages into a slack channel and automatically receive the transcription (including subtitle file) from the slackbot.

With AWS Rekognition, AWS helps aiconix train their models for facial recognition on a large scale. Parallel GPU power in the cloud gives them possibility to get results within a few hours, a process that previously required an entire day. aiconix´aingine has been online since February of 2019, and the prototype of the video recommendation tool, the SuperProducer, will be introduced in 2020.

With the combination of AWS Rekognition, other AI providers, and the company’s innovative features, aiconix offers customers a one-stop solution for AI features relevant to the media industry.