Using social intelligence to identify product concepts for the innovation pipeline

Case Study • NPD Concepts

Summary

A leading healthcare company looking to understand unmet needs and gaps in a growing and evolving child nutrition category in Vietnam markets leveraged Convotrack.ai's multimodal AI listening technology. Analyzing over 40K+ authentic consumer conversations and 8,000+ video content signals and cross-correlating with e-commerce performance metrics, our platform's neural decision engine identified nine high-potential product concepts across infant, toddler, and kids supplement segments. Two breakthrough innovations emerged with compelling ROI projections based on our 360-degree analysis.

Digital Landscape Analyzed

8300
Videos
💬
43K+
Conversations
🛍️
1MM+
Sales Volume

Details

Approach

Convotrack.ai deployed our proprietary “Mixture of Expert” LLM framework—a multimodal insight generation engine that combines unstructured data harvesting with semantic clustering algorithms. Our AI ingested multi-format data from 243 digital communities (24.4M unique digital identities), 8.1K+ video content assets (215M engagement touchpoints), 147 short-form videos (28M impression signals), and transactional data covering 20+ competitive entities with 500+ product SKUs.

The platform's voice analyzer mapped conversations from 415 key consumer personas, 35 expert nodes, and 106 influence amplifiers, allowing the AI to triangulate sentiment patterns across a combined network of 90M+ digital connection points:

Signal Acquisition Layer: Deployed natural language processing (NLP) crawlers to extract emerging pattern markers and trend vectors from unstructured data sources.
Neural Voice Analysis: Applied deep learning algorithms to isolate authentic consumer pain points and aspiration triggers across digital ecosystems.
Whitespace Mapping Engine: Utilized computer vision and language models to identify unaddressed format, ingredient, and messaging territories.
Market Sizing Algorithm: Deployed predictive modeling to forecast commercial potential through transaction pattern validation.
Concept Generation System: Leveraged generative AI to synthesize intelligence into actionable innovation blueprints.

Analysis

The platform's semantic clustering identified critical unmet need states through conversation pattern recognition:

For Infant Nutrition (0-12 months): Detected a 12.6% conversation density around organic ingredient preferences, revealing significant demand while validating competitive penetration with 73K+ monthly unit velocity worth 21.7M PHP ($370K).
For Infant Nutrition (0-12 months): Mapped consumer friction points around a specific enzyme deficiency (affecting 1:63 ratio of consumers) with no current solution architecture in-market.
For Infant Nutrition (0-12 months): Identified texture adaptation barriers during critical transition phases with 5.9% digital discussion volume.
For Toddler Nutrition (12-36 months): Isolated strong negative sentiment signals (8.7% conversation volume) regarding specific carbohydrate components with 76K monthly transaction frequency detected in competing products.
For Toddler Nutrition (12-36 months): Revealed personalization demand patterns with 41.1% of digital personas seeking modular nutrition solutions for evolving need states.
For Kids Supplements (36+ months): Extracted format preference data revealing 60.2% affinity toward specific delivery systems versus 33.9% for liquid formats.
For Kids Supplements (36+ months): Mapped sensory preference distribution showing 33.3%, 24.1%, and 17.9% signal strength for top three flavor territories.