Kensho Classify
Have a question about this Solution?
Kensho Classify is a flexible tool that offers three modalities for tagging your documents at scale, allowing you to choose the best solution for your topic tagging needs. With Classify, you can leverage our ‘zero-shot’ model for quick and custom tagging without training data, access our ‘out of the box’ concepts to tag your documents with relevant pretrained topics, or develop your own trained concepts tailored to your individual business needs.
- Determine if specific topics are present within a body of text and quickly navigate to where they occur.
- Annotate documents with tags and embedded metadata, making them richer, more searchable and discoverable.
- Drive business decisions with actionable market intelligence.
- Create data analysis dashboards using Classify metadata.
- Develop custom concept models with a low volume of data and without any machine learning expertise.
Service Provider Information
Kensho is the Artificial Intelligence accelerator for S&P Global that develops cutting-edge technologies to transform businesses. Kensho's elite engineering talent uses the latest advances in machine learning and the unparalleled breadth and depth of data available at S&P Global to create new actionable insights and solutions for decision makers.
Key Information
Use Cases
- Customize model for workflow-specific and niche concepts, allowing for flexibility and modularity
- Narrow in on specific topics. For example, company filings and earnings calls with ESG rating topics or legal documents with clause names
- Tag documents en masse and enable filtering and topic-related searches rather than simple keyword search
Benefits
- Fully Customizable: Dynamic training accessibility allows you to identify and tag domain-specific concepts, topics, and themes to meet niche workflow needs.
- Low-Lift Training: Unlike other text classification NLP tools, which often require large amounts of training data to produce a performant tagging model, Classify requires as few as 20 to 30 tagged paragraphs before tagging paragraphs automatically.
- Speed & Scalability: Ingest millions of characters of text at a time, doesn't require you to unnecessarily parse inputs to run effectively, and generates low-latency outputs.
- Time & Cost Effective: Unlike human tagging, which is time-consuming and expensive, or unsupervised AI data tagging methods that require huge amounts of data and ML engineering expertise, Classify serves as the Golden Mean for data tagging — requiring a minimal amount of human effort to produce results similar to large in-house ML teams.