Today, we're excited to introduce Keyword Hierarchies, a new way to visualize and understand the cascading nature of AI conversations.
Why we built keyword hierarchies
Traditional SEO taught us to think in terms of individual keywords: track "running shoes," and optimize for it. But AI conversations don't work that way and keywords must be tracked differently. When someone asks about "running shoes," they don't stop there. There is an inherent need to dig deeper. The conversation evolves and each layer reveals deeper intent:
- "running shoes" → "running shoes for women" → "running shoes for women with arch support" → "best running shoes for women with high arches under $150"
The technical innovation
The core engineering challenge was scale. With hundreds of millions of keywords in our database, we needed to make this massive dataset instantly searchable to build hierarchies in real-time. Our solution leverages advanced indexing and query optimization to traverse keyword relationships and calculate volumes on demand with sub-second latency.
Where we're headed
User queries in answer engines contain incredible richness, but this data is fundamentally unstructured. A single conversation about "marketing automation" might branch into pricing questions, feature comparisons, implementation guides, and integration concerns, each revealing different user needs and intentions.
We're exploring advanced techniques to better extract and organize these patterns and keyword hierarchies are just the beginning of this journey.
Get started by visiting the Conversation Explorer in your Profound dashboard or contact our sales team for a demo.
