Komprise unveils KAPPA to prep unstructured data for AI

Komprise has launched a serverless compute product for unstructured data, aiming to help organisations build custom data processing functions and run them across large file estates without managing infrastructure.

The product, Komprise AI Preparation & Process Automation data services (KAPPA), targets unstructured data stored across network-attached storage, cloud storage and software-as-a-service platforms. Komprise is positioning the release around growing demand for AI-ready data and the need to extract and enrich metadata at scale.

Unstructured data often sits in large volumes across multiple storage silos, with inconsistent naming, incomplete context and complicated access controls. These issues can hinder search, governance and downstream analytics workflows.

Metadata extraction and enrichment is a key step in preparing unstructured data for AI use. It can make content easier to discover and provide attributes used for security classification and policy enforcement.

Custom processing

Many IT teams still rely on established extract, transform and load approaches and pre-built connectors. These methods can be slow to build and expensive to maintain as requirements change. KAPPA is positioned as an alternative: teams write a per-file function in code, while the platform handles execution and scaling.

Users add a small amount of code in a data operation field, and the platform runs it across a specified dataset as part of a broader workflow. This approach is aimed at organisations that need specific metadata fields or tagging rules that vary by department, security policy or industry practice.

Examples include healthcare teams working with DICOM medical imaging files and reading custom headers for tagging. Workflows can also apply ERP project tags, mask personally identifiable information, import sensitivity labels and connect file context from other systems such as electronic lab notebooks. In media and entertainment, one example focuses on tagging files with selected EXIF metadata from digital media assets.

Komprise also links the service to agentic AI, where multiple AI tasks run in an autonomous sequence. It argues these systems require parallel execution across datasets and rapid invocation of data operations as new tasks arise.

Platform integration

KAPPA data services integrates with Komprise's existing products. Enriched metadata tags feed into the Komprise Global Metadatabase, an index layer for discovery across hybrid storage. Search and query run through Komprise Deep Analytics, while workflow orchestration uses Komprise Smart Data Workflows.

KAPPA also includes lifecycle management steps around a workflow, including automated pre- and post-processing. For example, it can start a cloud AI service before processing and shut it down once the job completes. Komprise and its partners are also developing a library of reusable data services that customers can configure for their environments.

Komprise presents KAPPA as "serverless compute" for unstructured data, with the core promise being operational abstraction. Teams focus on per-file logic, while the platform distributes work across large numbers of files and coordinates execution across storage locations.

"Enterprises are realizing that the unstructured data that has been piling up for decades is now a goldmine for AI, but it's incredibly hard to tap into," said Kumar K. Goswami, co-founder and CEO of Komprise. "Since nearly every enterprise has unique needs, KAPPA data services deliver a nimble, serverless compute architecture for custom metadata enrichment at scale."

Partner company P1 Technologies highlighted the need to preserve and restore context that can be lost when files move between systems or land in storage without their original metadata. It said the issue affects media workflows, research environments and records management, where context often sits in upstream tools.

"Valuable metadata including embedded information from media asset management tools and contextual enterprise-specific metadata which are key for AI are lost when files are stored," said Aaron Cardenas, CEO of P1 Technologies. "Kappa data services allows us to enrich file metadata and customize it to our clients' needs with incredible ease and speed, while also conforming to their security models. This new functionality from Komprise is making a tremendous difference in the outcomes of our projects."

KAPPA data services are available through an early access programme.