Sentra Launches Breakthrough Unstructured Data AI-Classification Capabilities, Providing What Cyera and Other DSPMs Can’t - 99% Accuracy, Petabyte Scale at a Much Lower Compute Cost
The DSPM leader’s new specialized Small Language Models (SLMs) empower global organizations to adopt GenAI faster while conquering the massive risk of unstructured data
Sentra announced a new AI Classifier for Unstructured Data capability, breaking through previous technology barriers to solve one of the most critical risks facing enterprises. The new design incorporates specialized Small Language Models (SLMs) within Sentra’s proprietary AI classification engine, empowering organizations to adopt generative AI (GenAI) faster and avert sensitive data exposures hidden within unstructured data.
Unstructured data, representing more than 80% of enterprise data - poses the greatest risk for catastrophic “data extinction” incidents, such as critical data disclosure or breaches. Sensitive data and secrets often lie dormant, waiting to be exploited if not discovered in time, yet most organizations remain blind to the risk. Current DSPM solutions, including Cyera, BigID, Varonis and others, fail to solve this due to limitations in accuracy, efficiency and scalability, making a new approach essential.
“What impressed us most is how Sentra's AI classification engine continuously learns from our environment. It not only identifies sensitive financial data at scale but evolves with our business - giving us confidence to safely integrate GenAI into more of our workflows," said Zachary Schulze, Sr. Staff Application Security Engineer, SoFi.
Sentra’s new capabilities allow enterprises to classify petabytes of unstructured data at scale with unprecedented accuracy and efficiency, helping security teams safeguard critical information, prevent sensitive data exposure and accelerate AI-driven innovation. While existing vendors rely on simple pattern matching algorithms or generalized, compute-heavy LLMs, Sentra’s domain-specific AI SLMs fully understand the context around unstructured data to accurately detect and label critical data.
Uniquely Determines the Full Context of Enterprise Data
Unlike traditional DSPMs, Sentra’s AI Classifier engine deciphers the sensitive and proprietary data unique to each organization. Each SLM-based classifier is ‘domain optimized’ to surface the full context of unstructured assets and files by categorizing departments (such as finance, legal, HR and marketing, etc.), industry (healthcare, retail, etc.), geography, ownership (employee, customer, etc.), sensitivity and company-specific intellectual property. This segmentation ensures classification accuracy far beyond regex or metadata-based approaches which only identify a match, not meaning. Sentra AI unstructured data classifiers deliver context with superior accuracy (>99%) and high confidence, facilitating automation and streamlined data governance.
Continuously Learning AI
Sentra AI Classifiers continually learn with every scan, identifying IP and proprietary, business-specific data unique to each customer and industry. After an initial scan, the platform automatically learns to identify proprietary and business-specific data categories without requiring rescans.
This is achieved by understanding the semantic context of all documents in a customer’s environment without copying the data and then clustering the different documents by their type, using advanced and continuously learning AI models. For example, the engine can progress from recognizing common data types, such as tax documents or invoices, to pinpointing highly specialized categories like pharmaceutical drug research, beverage and food formulas, or algorithm designs for special projects. Tagging data with this level of precision facilitates confident enforcement actions, eliminating the need for manual validation steps and guesswork. Sentra’s AI Classifiers support over 70 languages, with the ability to learn additional over time.
Efficiency at Scale with Small Language Models
Sentra’s compact, efficient SLMs drastically increase performance and reduce compute demand by an order of magnitude compared to LLMs deployed by other DSPMs. As the Sentra engine evolves, customers see continuous classification improvements and more granular classifications without added costs or rescans, which only occur when the underlying data changes. These innovations enable organizations to analyze hundreds of millions of documents cost-effectively, in stark contrast to the inefficiencies of applying general-purpose LLMs to classification tasks. Sentra achieves exceptional accuracy at a fraction of the compute cost, maintaining performance at petabyte scale where other solutions often degrade.
“Generative AI offers incredible potential, but it also introduces new data risks if sensitive or proprietary information isn’t properly governed,” said Yair Cohen, VP, Product and Co-Founder of Sentra. “Our AI classification engine provides unmatched scale, accuracy, context and adaptability - empowering enterprises to innovate with GenAI safely, responsibly and cost-effectively.”
Learn more about Sentra’s AI classification capabilities.
