CASE STUDY: CUSTOM AI-POWERED VIDEO DATABASE
August 7, 2023 | Case StudyThe Challenge
Our client, an industry-leading political research firm, needed a software application to help them organize their video assets, derive data-driven insights, and distribute the analysis and underlying video to clients with rigorous permission and security requirements.
Their pre-existing solution involved combining expensive, off-the-shelf software with significant manpower and still did not give them the control or client experience they desired. The client engaged Bullpen Technology Group to build an enterprise, custom solution incorporating the latest advances in artificial intelligence.
The Business Opportunity
The client’s business process relied on human analysts to transcribe, annotate, and distribute proprietary video assets to their customers. This time intensive workflow slowed down the time between a video’s creation and its use by a customer. Additionally, because they were repurposing other consumer-grade software, there was limited visibility into how (or if) the client’s customers were accessing the data.
Further, neither the client’s analysts nor their customers had the ability to search across the video library for keywords or attributes, including transcripts, speakers, or locations.
The Solution
Bullpen Technology Group worked with our client to develop a custom web application that automatically transcribed and analyzed their video assets using the latest artificial intelligence (AI) technology. Once processed, analysts could more quickly annotate and distribute the videos. Customers are notified about new, relevant videos and users have the ability to search the entire catalog of content based on titles, speakers, keywords, transcripts, and more. Further, our client now has total control over their business process giving them robust analytics into how their customers are engaging with video library.
A key aspect of our approach to a unique project like this is working with the client to truly understand their industry, their customers, and the problems they seek to address. This process necessarily precedes any discussion of software requirements to ensure Bullpen delivers to clients a solution rather than simply a product that meets specifications.
The Results
After a successful limited deployment for select customers mid-2022, our client introduced the software to all customers in 2023. Key results metrics:
- Year to date the application has processed more than 3,000 videos and feedback from their customers is overwhelmingly positive.
- Our client’s custom software application has improved the delivery time of new videos to customers by almost 40% over the previous year.
- With their enhanced analytics, the client can monitor engagement to see that users, on average, log-in to the platform 3.5 times every week.
Conclusion
Bullpen’s approach to diving deep on our client’s existing process, studying the challenges they faced, and identifying the right tech stack was essential to deliver a product that truly addressed their needs.