IT teams dread the risks of sensitive data lurking where it shouldn’t be, risks that are compounding as unstructured data grows explosively across all industries.
Generative AI has made the situation more troublesome. Attempts to input PII (personal identifiable information) into GenAI platforms represent over half (55%) of data loss prevention (DLP) events, followed by confidential documents (40%), according to 2024 research by Menlo Security.
The global average cost of a data breach reached $4.88m in 2024, according to IBM. Increasingly, storage administrators are responsible for data governance and compliance but lack ways to do this systematically across their data estate.
To help mitigate the problems, and offer a new sales opportunity for its partners, unstructured data management specialist Komprise has unveiled sensitive data detection and mitigation capabilities to help organisations prevent the leakage of PII and other sensitive data to AI, and reduce the risk of potentially ruinous data breaches.
Komprise’s Smart Data Workflow Manager now includes detection for PII, and regular expressions and keywords, to simplify and automate the process of finding and tagging sensitive data and moving it to protected locations. The additional capabilities were promoted at last week’s IT Press Tour of Silicon Valley, attended by IT Europa.
The new sensitive data management capabilities of Komprise Smart Data Workflow Manager include:
Standard PII detection: Select which PII data types to scan, such as national IDs, credit card numbers and email addresses. Komprise supports multiple classifications to identify multiple types of PII within any given file.
Custom Sensitive Data Detection: Customers can find any text patterns in their data via both keyword and regular expressions (regex) search to identify specific data formats like employee IDs, machine or instrument IDs, product or project codes, or even PHI data like healthcare-system specific patient record IDs.
Scans Sensitive Data in Place: Executes locally behind enterprise firewalls so sensitive data stays in place, unlike cloud-based data detection services.
Remediate and Move: Once Komprise identifies sensitive data, users can set up a workflow to take appropriate action, such as confining the data or moving the data to a safe location.
Pre-Process for AI Ingest: Sensitive data detection can be a pre-process step for an AI ingest workflow to eliminate sensitive data leakage to AI.
Ongoing Workflow: Users can set workflows to run periodically, so Komprise automatically finds and acts on any new sensitive data for ongoing detection, tagging and mitigation, with full audit capabilities.
Common use cases for the new sensitive data detection capabilities include:
Prevent unintended data leakage during AI Ingest: Data governance for AI remains a top priority. Komprise automates the workflow of identifying and excluding sensitive data from the data copied to AI, thus preventing human error and risky data leakage during AI ingestion.
Sensitive Data Handling for Cyber-Resilience: Roughly 80% of data breaches involve sensitive data, according to Verizon’s Data Breach Investigations report and other sources. This is especially risky in regulated industries such as healthcare and finance where organisations face hefty penalties. By automating the process of finding sensitive data in non-compliant places and proactively remediating by moving this data to secure storage, organisations can reduce the risk of sensitive data breaches.
AI Data Workflow Auditing: Komprise maintains a full audit record of all data processed by any workflow, such as a workflow copying data to a location for ingestion by an AI or ML system.
“The risk of sensitive data breaches is escalating and thereby paralysing organisations from using AI,” said Kumar Goswami, CEO of Komprise. “We are systematically reducing sensitive data risks so that our customers can improve their cybersecurity and provide data governance for AI ingestion with the new sensitive data detection and mitigation capabilities in Komprise Smart Data Workflow Manager.”
