Created an AI-powered solution that automated 100% of video call auditing for financial institutions, transforming a manual 5% sampling process into comprehensive coverage with deep analytical capabilities. The solution later expanded beyond video to phone operations, creating an entirely new revenue stream.
Role
Product Designer
Industry
Banking / FinTech
Duration
26 weeks
Team Size
5 members
Status
Launched & Scaling
Context
Leading the product strategy at a video technology company during the AI boom, we faced pressure to integrate AI meaningfully rather than superficially. Our clients, primarily financial institutions, were manually auditing only 5% of their video calls due to volume constraints - processing 15,000 daily calls (75,000 weekly) required significant human resources. This limited coverage posed risks and missed opportunities for operational insights.
When AI became accessible technology, I focused not on adding AI features for their own sake, but on finding capabilities that could meaningfully transform how our clients operated. This led to examining core operational challenges that could be solved through AI automation.
Opportunity
Analysis revealed that clients spent enormous resources to audit just 5% of video calls, taking two full-time auditors to review this small sample. With 15,000 daily calls averaging 5 minutes each, the manual process couldn't scale. This presented a perfect opportunity for AI automation to achieve 100% coverage while dramatically improving analytical capabilities.
The existing process was clearly insufficient - imagine auditing only 5% of a company's financial transactions or 5% of a car's safety features. But the manual workload made it impossible to scale: 75,000 weekly calls equated to 6,250 hours of content. Even with full-time auditors, comprehensive coverage was impossible without automation.
Solution
Developed an API-first solution using ChatGPT API and custom audio processing that:
Extracts and enhances audio from video calls
Processes speech through ChatGPT API for accurate speaker-separated transcription
Provides automated analysis including call ratings, trouble word detection, and tag verification
Delivers insights through an intuitive interface with email-like notifications
Expands beyond video to phone operations through API flexibility
The API-first approach proved crucial, enabling expansion into phone operations and creating new revenue streams from previously unreachable market segments. This strategic architecture decision transformed a video-specific tool into a comprehensive communication analysis platform.
Results
Drove 40% of company revenue through new AI-powered solution
Achieved 87% client adoption rate as an upsell product
Generated estimated $1.5M annual revenue through API expansion
Transformed 5% manual audit coverage to 100% automated coverage
Created new market opportunities beyond core video business
Key Takeaways
API-first design enables unexpected business opportunities
Focus on capabilities over features drives innovation
Strong foundational architecture enables rapid market expansion
Solving core problems thoroughly opens adjacent market opportunities
