Case Study: Empowering Next-Generation Adhesive Innovations with a Hybrid Patent Search
When a cutting-edge adhesives company like Biobond wants to explore patents in their field—such as advanced bonding solutions, new polymer formulations, or eco-friendly adhesives—they need more than a simple keyword-based approach. Our self-service patent search platform combines both traditional keyword matching and semantic understanding, delivering fast, pinpoint results. Here’s how it works, condensed into four key stages.
1. Defining the Search & Generating Keywords
The process starts when you provide a detailed description of what you’re looking for—maybe it’s advanced bonding agents, innovative polymer formulations, or eco-friendly adhesives. For a company like Biobond, that might mean entering terms such as “bio-based polymer adhesives” or “rapid-curing adhesive solutions.”
Behind the scenes, an advanced language model analyzes your description to pull out the most important terms. You remain in control—adding, removing, or refining these keywords to capture the technology scope precisely. This way, if you already know you’re looking for a specific chemical formulation or a particular bonding technique, you can guide the system to focus on those details.
2. Capturing Semantic Context
While the keywords are being generated, your description is also converted into a numerical representation, often called an “embedding.” This captures the deeper, contextual meaning of your text, so the system can understand nuances like “low-VOC adhesives” or “biodegradable bonding agents,” even if patents use slightly different wording.
At this stage, you can choose filters such as assignee (e.g., specific companies), date range (for instance, only patents from the last 20 years), classification codes (CPC), or specific inventor names. This ensures that if Biobond only wants to see patents filed after a certain year or from a particular subset of the adhesives industry, the system narrows down those results accordingly.
3. Executing a Dual-Layer Search
With your refined keywords and any filters in place, our platform first searches a dedicated database of granted U.S. patents from the last two decades. It quickly gathers documents that explicitly mention your selected terms—an immediate snapshot of potentially relevant prior art.
Next comes the second layer: semantic search. Each patent in our 5M+ dataset is already stored as an embedding. We compare your query’s embedding against these to see which patents share conceptual similarities with your text. Even if a patent describes the same innovation using different words (“thermally resistant adhesive” instead of “heat-tolerant bonding agent,” for example), it will still appear in the results because of its high semantic match.
4. Reviewing the Top 100 & On-Demand Insights
After combining both keyword relevance and semantic similarity, the platform presents you with the top 100 patents. This list provides a solid starting point for deeper investigation—especially helpful for companies like Biobond that need a broad yet targeted overview of adhesives-related patents.
Finally, if you want deeper insights on any specific patent, you can simply create a new column in your results and pose a follow-up query. The system will analyze and summarize each patent for you, answering questions like “How does this patent approach polymer-blend formulations?” or “What performance improvements are claimed for bond strength?” This streamlined, interactive feedback loop helps companies quickly evaluate which references are truly relevant to their R&D objectives.
Why It Matters
For a company working at the forefront of adhesives technology, the ability to uncover both explicitly mentioned and contextually relevant patents is invaluable. By blending traditional keyword matching with advanced semantic embedding, our platform ensures you won’t miss patents simply because they use different language. And with AI-driven summaries at your fingertips, you can spend your time innovating—rather than sifting through mountains of text.
Whether you’re formulating next-generation adhesives or exploring other bonding solutions, this powerful hybrid search tool provides a holistic view of the intellectual property landscape, empowering you to make informed decisions, refine your strategies, and drive innovation forward.
