close
close

Komprise Smart Data Workflows bring AI models to all your data

Unstructured data management company Komprise has added a new Smart Data Workflow Manager feature to make automated data augmentation easier.

Komprise already works with large-scale unstructured data to provide analytics and reporting on how data is used and where it is located. Global search allows companies to locate data across their entire estate, regardless of storage provider. Data can be moved from fast, expensive storage to cheaper archival storage based on enterprise-wide policies.

But collecting large amounts of data is of little use unless you can find what you’re looking for when you need it later. Finding what you’re looking for means understanding what and where it is. Without a well-constructed index, this task is virtually impossible.

“Ultimately, companies store unstructured data so they can extract value from it. So their users can easily find the data they need, no matter where the data is located,” said Krishna Subramanian, co-founder, president and COO of Komprise.

With Smart Data Workflows, Komprise can rely on external services to tag and index disorganized piles of corporate data based on their content. Image recognition algorithms can be run on photos or videos, natural language systems can process text, and so on. Each additional workflow complements the existing Komprise global index to provide some more structure to the unstructured data tsunami.

Komprise serves as a data librarian for the company, maintaining the collection of old books, photographs and recordings of obscure speeches, noting who was mentioned and the topics they discussed. By connecting Komprise to external tools with workflows, it can automate the otherwise endlessly tedious process of manually reviewing the contents of files to figure out what they’re about and writing them down.

Even the best reference librarian can’t read, watch, or listen to everything, especially with the amount of new data created within an enterprise every day. No human worker has the iron will and fierce determination required to flag down every mention of the word lever across the company’s entire SharePoint domain. And then again, looking for pictures of the mark 3 turbo encabulator.

Komprise can be connected to external analysis tools specifically tuned to recognize Mark 3 turbo encabulators or the sound of the Wilhelm scream. These scanners work in parallel and can extend an enterprise dataset with tags that mark the data for easy retrieval. Komprise can limit the amount of data fed into these sometimes expensive analytics tools, keeping costs under control and focusing on high-quality data sets as a priority. Once tagged, these datasets can serve as input for even more analyses, working in parallel to provide a rich set of indexes to an enterprise data library.

“It’s not limited to one cloud, or one account, or one storage system, because we keep it in a central place everywhere,” says Subramanian.

Due to its global perspective, Komprise can provide a shared data librarian service to the entire organization. Isolated storage dumps can be revived as source repositories for new processes, combining with other knowledge bases to provide a more holistic view.

The librarians at Komprise customer Duquesne University have done just this to process the university’s digital collections. To create a retrospective on John F Kennedy and his assassination, the university used a Komprise workflow linked to a cloud image recognition service to process the images.

“Instead of spending 20,000 minutes searching through 10,000 images to try to find the images that had JFK, they ran this workflow and in less than two hours they had sent the data to the cloud, got the answers and nine matches found it. Subramanian said.

Komprise does not aim to replace specialized tools such as sensitive data detectors or legal contract language parsers. Instead, Komprise wants to give customers an easy way to connect these specialized tools to business data in an organized, repeatable way. Once configured, the Smart Data Workflows can be executed automatically so that new data is processed correctly. Each new wealth of raw materials can be quickly added to the knowledge corpus using the latest and best techniques.

As a fan of libraries and well-organized metadata, this new feature from Komprise is very interesting. Few companies have the foresight or discipline to hire librarians to organize their information, which is strange considering we live in the information age. AI systems can also be fraught with dangers and mislabel and archive things, but on a massive scale. Yet this is a task that is often not performed at all. Companies store vast amounts of data, much of which may be worthless. to be sure. The waste is enormous.

I think it would be a good thing to create a little more order in the rapidly escalating chaos.