Visualyst uses AI to crawl videos for hate speech, violence, and other compliance issues

About Visualyst

  • Founders: Miguel Silva and Mustafa Adam
  • Founded: 2019
  • Employees: 6
  • Amount raised: €560,000
  • Ultimate Goal: To Help Businesses Efficiently Process Video for Safe Viewing

Reviewing the audio and visual content of a video to ensure that it complies with a country or platform’s regulatory compliance is a long and complex process. It’s a task usually done manually by small armies of editorial teams and takes a lot of time and repeated oversight to get it right. In this episode of Start-up Of The Day, Founder Miguel Silva tells Innovation Origins how Visualyst stepped in to help these teams get to work faster than ever.

What can Visualyst do?

The startup aims to increase efficiency in the video delivery and streaming space. Or really any type of video distribution platform. More and more videos are produced around the world and are intended for distribution and consumption around the world. Usually it has to go through some sort of moderation where someone reviews the video and decides if it’s breaking some type of regulation. This takes lots of time. We have digitized this process and made it very efficient. We use technologies such as artificial intelligence (AI) to explore the video and detect elements that may be problematic. Then it brings it to the attention of human moderators, allowing them to make decisions faster. We save organizations a lot of time and money.

What is the benefit of using Visualyst?

It depends on the complexity of the video, but Visualyst takes between 10 and 30 minutes to crawl every hour of video. Organizations typically have teams that review video, spending between 2 and 5 hours reviewing each hour of video programming. More importantly, everything after that is also much faster because we offer collaboration in the cloud. Human moderators can very quickly collaborate and assign each other different events, tag each other and include comments in a Google Doc collaboration style. They can then make very quick decisions about whether a piece of content is ready to stream or not. Visualyst helps teams get to market faster with content. This eliminates the very long backlog of videos they have to review, which they do today. There is also a growing need for organizations to keep up to date not only with regulations, but also with their editorial aspect. This begs the question: what kind of content do you want to associate with your brand?


Does Visualyst work?

We’re not looking to replace review teams, but we’re looking to give them a great tool to be much more effective. And if you can bring 35-50% efficiency to the process, which is what we’re aiming for, you’re potentially saving some companies half a million dollars a year. We kind of have a transactional cost. We charge you per minute of video we review or per hour of video we review. So it depends a lot on the volumes you have. If you have small volumes, you pay less than if you have, you know, hundreds of thousands of hours of programming that you need to review. We needed to find a way to scale and be relevant for very small business users all the way up to very large users. As part of an EU-funded project, we have also developed a model that specifically searches for personal information in a video. Another thing we’ve developed is the ability to find any type of background text that is insulting or hate speech – think graffiti that may be on the wall in the background or T-shirts that may have a hate symbol. These are things that are very easy for a compliance officer or publisher to miss. In these projects, we can actually increase the level of efficiency of AI to 35% more than that of the human eye.

How to use Visualyst?

The product is already commercially available. As a user, you log in to and create an account. You can then create a workspace to which you can also invite collaborators. In this workspace, you can create a digital policy, for example by reviewing videos for child-friendliness. Such a child-friendly policy will detect instances of alcohol, excessive violence, hate symbols, etc. Thus, you select from a large menu of instances to report and include in your strategy. Today you need to create your own policies, but over time we will create a repository of predefined policies. The next step is to upload the video. Visualyst then analyzes the video according to your policy. This process can also be automated so that you can have the system retrieve your videos every Friday, analyze them according to this policy, and prepare the results for my team to leave on Monday morning. From there, the team sees a nice interface where they can play the video and see an overview of each detection encountered by category. We give them a timeline that they can scroll back and forth. Once they’re done, they click a button and export the data to a video editing tool so they can make any necessary edits. A report is also produced. We think the software should be very intuitive. If your cloud software users need a user manual these days, we think you’ve missed the mark.

What idea sparked the creation of Visualyst?

The founders are people with a background in broadcasting. I worked for Norway’s largest commercial broadcaster. It’s a very innovative broadcaster that usually develops a lot of technology and I was selling some of those platforms to other broadcasters around the world. I’ve certainly spoken to many different broadcasters over the past 10 years and started to realize that the technology was the easiest part to solve. The biggest challenge was always waiting for the content to be approved. And we started to dig a little bit on that. I’m sure people notice that movies come out one day in the United States and then several months later they see it in another country. It has a lot to do with compliance and backlogs which are quite problematic in the industry.

John C. Dent