Course5 VideoSense

As human beings, we love consuming content in a Video Format, which is by far the most popular engagement medium in the digital landscape. Here are some interesting facts for your consideration:

• For 66% of the people, video is their primary source of information.
• Nearly 500 million people watch Facebook videos daily.
• More than 90% of all video views come from mobile devices.
• YouTubers upload 500 hours of video content on their platform every minute.
• 93% of the businesses gain new customers due to branded video content.

At Course5, we work with some of the most eminent names in the Fortune 500 enterprise list, and a significant % of our clients invest millions of dollars annually in creating branded content, facilitating panel discussions, concept videos, research studies, and experimentation. We have also noticed hundreds and thousands of videos created by influencers annually, which refer to our enterprise customers' brands, categories, and product names.

With this plethora of content created annually, enterprises need help finding actionable elements in these conversations to improve their customer experience, market share, supply chain, revenue, and overall profitability.

At Course5, we have built a unique accelerator called “Video Sense” for today’s digital enterprises. This accelerator leverages Computer Vision and the power of LLMs to demystify the hidden insights in video conversations and express them intuitively and succinctly to various personas. Each insight gets auto-evaluated from the relevance, coherence, coverage, and neutrality points of view before it gets published via the User Interface or pushed as a response to an API request.

The accelerator has been built leveraging the Microsoft Azure Stack. It uses Azure Cognitive Services, SharePoint DAM capabilities, and Azure DevOps best to deploy the accelerator as a service in the client’s Azure cloud infrastructure and expose it via API endpoints. We also deliver a bespoke Gen-AI experience layer leveraging the underlying OpenAI GPT model behind the scenes. We reveal the UI via SharePoint, PowerApps, or a custom React and Angular-based interface.

The accelerator beautifully identifies the speakers in the discussion, their major themes, what are the key takeaways the consumers of these insights can derive, and, most importantly, the opportunity to leverage the power of prompt engineering to dig deeper into the generated insights and quench the thirst of the varied personas.

The insights can be consumed as a raw transcript, an executive summary with critical topics, a knowledge graph reflecting the inter-relationship among different topics and speakers, topic-wise segregation of video clips for replay, and actionable takeaways for the business user consuming the insights. Here is the metadata our models extract from the Videos, enabling you to store in any enterprise DAM platform (e.g., Adobe Experience Manager, Canto, Cloudinary, Bynder, etc.)

• Speaker (Videos Only)
• Transcripts
• Topics
• Keywords
• Speaker Diarization
• Speaker Recognition
• SentimentsObject & Scene
• Speaker Face
• Text Extraction (Subtitles, Slides, Whiteboards, etc.)
• Emotions
• Culture Insights
• Demographic Insights
• Segmented Scenes
• Activities with-in Video

Soon, we will be releasing the following capabilities:

• Parallel pre-processing and summarization of 100s of videos at scale.
• Aggregated summarization of interrelated videos by topic.
• Split the Input Videos to multiple brand/topic/influencer videos.
• Determine the overall sentiment of the discussion, speaker, and the discussed topic.
• Identification of key influencers in the discussion.
• Generating targeted campaign videos for varied channels by adhering to their content publishing guidelines. (For e.g., Tik Tok, YouTube, Instagram, Facebook etc.)

To learn more about how Video Sense accelerator, along with the Azure Cognitive AI stack, can bring value to your business, request a 2-week consulting engagement and explore the possibilities of doing a Proof-of-Concept for six weeks leveraging your enterprise data within your secured and scalable Azure data infrastructure.