Patent Trend Analysis using LLMs for EV Batteries (EU Enterprise)
Case Study Summary
Client: Confidential / Leading European EV Battery Manufacturer
Website: —
Industry: Automotive (EV batteries)
Impact Metrics:
- −90% review/labeling time per weekly batch (from 5 days → 0.5 day)
- −90% manual classification overhead for the IP team
- ≈ €490k/year cost savings (10 IP engineers @ €50k/yr; weekly task reduced 5d → 0.5d → 45 person-days saved/week × 48 weeks × €227/day)
- Analysts reallocated from manual tagging to higher-value trend interpretation
The IP team needed faster, consistent insight into thousands of new patents to inform R&D on emerging EV-battery technologies. We delivered a pipeline that turns raw patent exports into titled clusters with summaries and a trends dashboard.
Challenge
IP engineers were manually scanning and tagging multilingual patents—slow, inconsistent, and hard to replicate at scale. Multilingual content and unstructured abstracts made it difficult to compare filings and report consolidated trends to innovation stakeholders.
Our Approach
We built REST endpoints for language detection → translation → domain-specific embedding → unsupervised clustering → cluster summarization → trend tracking, exposed via a FastAPI service on Azure and backed by a vector index for semantic lookups. Clusters receive concise titles/summaries and can optionally align to IPC categories for consistent reporting.
Results & Impact
- Weekly patent exports processed into digestible, titled clusters with auto-summaries
- Review loops cut from hours to minutes per batch
- Earlier visibility of emerging themes; analysts focus on interpretation vs manual tagging
- Stable, repeatable releases with containerized CI/CD
Solution Overview
Baseline EV-battery patent trend analytics solution architecture
Tech Stack
- Cloud: Microsoft Azure Cloud Infrastructure
- CI/CD: Azure DevOps Pipelines
- Containerization: Docker
- Data Platform: Azure Databricks (APIs for ingestion & jobs)
- Vector Index: Databricks Vector Search
- Backend: Python services with FastAPI (REST)
- Language Detection: XLM-Roberta
- Translation: mBART-large-50 many to one
- Embeddings: BERT for Patents (fine-tuned for EV-battery patents)
- LLM (summarization): Mistral 7B
Additional Context
- Timeline: ~4 months
- Team Size: 4 people (2 Data Scientists, 1 MLOps Engineer, 1 AI Tech Lead)
- Role: AI Tech Lead
- Collaboration: Close with IP analysts for evaluation loops and taxonomy alignment
- Future Plans: Feedback signals into clustering/evals; optional supervised topic labels
Three Key Points
- Domain-specific embeddings tuned for EV-battery patents to improve cluster cohesion.
- Unsupervised clustering with optional IPC-guided labels for standardized taxonomy.
- Trend analytics dashboard to monitor cluster momentum and surface emerging themes.
-
Let's have a virtual coffee together!
Want to see if we're a match? Let's have a chat and find out. Schedule a free 30-minute strategy session to discuss your AI challenges and explore how we can work together.