Komprise launches KAPPA to hunt metadata across enterprise file silos

Komprise's KAPPA service automates the collection of metadata tags by AI agents from unstructured data distributed across many silos.

The company envisages a situation where an organization's unstructured data is spread across multiple storage silos – filers, cloud stores, and SaaS services – with no overall namespace. It says that data selection for AI purposes is hindered by the lack of a central metadata repository describing these distributed unstructured data sources. Its aim with KAPPA (Komprise AI Preparation & Process Automation) is to remedy that by enabling customers to set up their own custom metadata enrichment workflows by specifying and invoking KAPPA functions.

Kumar Goswami, co-founder and CEO of Komprise, said: "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. Since nearly every enterprise has unique needs, KAPPA data services deliver a nimble, serverless compute architecture for custom metadata enrichment at scale."

KAPPA functions obtain metadata tags from specified datasets and load it into the Komprise Global Metadatabase Service. This makes supported hybrid storage data sources searchable and discoverable. Komprise's Deep Analytics software can provide such a search and discovery service, and feed the resulting data to orchestrated AI workflows using Komprise's Smart Data Workflows functionality.

Komprise will automatically manage specified pre- and post-processing for the KAPPA workflow, such as spinning up a cloud AI service prior to processing and decommissioning it when the data service is complete. KAPPA is a serverless offering with Komprise handling "the complexities of scaling and executing the instructions across large data sets and infrastructure" – large as in petabytes.

These KAPPA functions can be used and invoked by agents in agentic AI environments. As an example, an airline's customer service AI agent could tag all files for a journey with its reservation number using a KAPPA function.

Another Komprise example cites a research director in a healthcare organization who may want to read custom metadata headers from medical DICOM files for tagging, apply ERP project tags to files, mask personally identifiable information, or import sensitive data labels, and integrate project context from other platforms such as Electronic Lab Notebooks (ELNs).

A KAPPA function is set up in the Komprise environment by IT and data experts, who, using KAPPA data services software, insert a few lines of Python code for the requested actions per file into a data operation field. The software "then performs the steps to execute the custom action across a specified dataset as part of a broader AI workflow or data management plan."

Komprise and its partners are developing a library of reusable data services that users can configure for their specific requirements. KAPPA data services are currently in an early access program for customers. Learn more here.

Bootnote

A Kappa also happens to be a green, scaly human-like water monster - sometimes more dangerous, sometimes depicted as simply a mischievous sprite - living in ponds and rivers in Japanese folklore.