Patent searching: Concept Tree

Patent searching: Concept tree

See your search, trust your results

Patent searching to find relevant records is now faster by visualising technology as connected concepts, in a ‘concept tree’, not disconnected keywords

We’ve heard you. Patent searching shouldn’t feel like solving a puzzle every time you need to find prior art. For too long, IP professionals have wrestled with complex Boolean queries, spending hours refining search strings only to end up with either 50,000 irrelevant results or missing critical patents entirely.

That’s why we’ve added Concept tree to Minesoft Origin 2, a feature that transforms how you search for patents by turning natural language descriptions into interactive, visual graphs. Instead of translating your invention into disconnected keywords, you’ll see and manipulate the actual technical structure of what you’re searching for.

The result? Faster, smarter, and more intuitive patent searches that align with how you actually think about technology.

What is Concept Tree?

Concept Tree allows you to input a text description or patent number, which is automatically converted into a dynamic knowledge graph. This graph visualises the technical components and their relationships, as connected nodes you can interact with, refine, and explore.dsgsd

Unlike traditional keyword searches that treat terms as isolated fragments, Concept Tree understands that inventions are systems of interacting components, functions, and outcomes.

It mirrors the way IP professionals and inventors naturally think about innovations: as structured relationships, not random word combinations.

Why it matters:

Today’s patent searchers face a frustrating reality. A simple keyword search for “smart home security system” might return 50,000 patents, most of which are irrelevant. Constructing a precise Boolean query like (AI OR "artificial intelligence") AND "smart home" AND ("threat detection" OR surveillance) this requires deep expertise, trial-and-error, and still risks missing results if the terminology doesn’t match exactly.

Even standard AI search presents its own challenge: you don’t know what exactly the AI will focus on in your paragraph, creating a black box where the search logic remains invisible and unexplainable. Learn more about other AI capabilities.

Concept Tree eliminates this friction by automatically structuring your search logic visually and semantically, finding patents based on technical relationships and conceptual similarity, not just word overlap.

Perfect for: patent professionals who deal with complex technologies but aren’t necessarily search syntax experts and those who want to eliminate the AI black box.

Concept Tree changes everything in Patent Searching

Find what Boolean queries miss: surface patents that are conceptually related but use different terminology.

Visual, intuitive refinement: see your search logic laid out as a graph: “Security System → Camera Subsystem → AI Processing Unit → Threat Detection Algorithm.” Want to broaden results? Remove a node. Want to focus on edge computing? Add it. Every change is visual and immediate.

Faster, more focused results: instead of manually reviewing 5,000+ mixed-relevance results, Concept Tree returns highly relevant patents ranked by structural and semantic similarity.

Cross-domain discovery made easy: innovations often combine ideas from multiple fields. Now, you can naturally integrate concepts from various domains into a unified query.

Ready to transform your search workflow?

Unlike black-box AI searches, where the logic remains hidden, Concept Tree makes your search reasoning transparent and explainable, giving you full visibility and control over what the system focuses on.

Interested in learning more about patent searching & concept tree?

Join our webinar! Register here.