Innovative knowledge-sharing tools elevate the modern workplace

Data discovery

The conversational AI implications of Chevala’s view are both apparent and relevant to the natural language technology facets of search. The data discovery applications of such capabilities are equally as important, if not more so, when one considers where organizations are searching. There’s certainly a case to be made for uploading enterprise knowledge into an engagement platform dedicated to this application.

This method supports multiple formats, “whether it’s video, a PowerPoint, Google Doc, Microsoft Word, or PDF—no matter what it is,” Greenberger said. “You can mix types of content in one post or create stories where you have a collection of related pieces you want in front of others.” In such centralized settings, certain solutions facilitate “rule-driven discovery of asset types, i.e., personally identifiable information,” Wills said. Organizations can also search through individual platforms via a centralized knowledge management solution. The use cases this approach supports are foundational to knowledge sharing and include:

♦ Overcoming silos: Traditional silo limitations mean users don’t know what’s in each silo. Accomplished search capabilities reveal the contents of different platforms, “be it SharePoint, or Confluence, or Dropbox, or Jira, GitHub, Slack, or email,” Hoeffel said. “These are all giant knowledge repositories that conceal from the searching public the data you need to do your job.”

♦ Locating additional relevant content: By enabling users to search across different platforms, formats, schema, and file types, data discovery search applications surface results users were not aware of—but that are equally valuable to their searches. “You can find ‘the’ document, but more than that, you can find related documents you didn’t know existed that are similar to the one that you know you’re looking for,” Hoeffel said.

♦ New employees: One of the foremost deployments for compiling, curating, and maintaining enterprise knowledge is the capacity to readily position it in front of new hires that need it to do their jobs. “This discovery capability is of huge value, especially to those new to organizations that don’t know all this stuff was there,” Hoeffel said.

♦ Mergers and acquisitions: Such discovery capabilities are vital for organizations that merge with or get acquired by others and suddenly have different databases and repositories to search. It’s also critical for ascertaining if an acquisition prospect has knowledge that’s worth acquiring.

Natural language technologies

There are several aspects of natural language technologies that are optimal for knowledge sharing. Taxonomies, for example, are an excellent starting point for text analytics. “You can use them for entity extraction,” Aasman said. “You can do a process where you use the taxonomy and give all the names and the synonyms to the entity extractor, and it will extract, for example, every router that it finds in any text.” Statistical cognitive computing approaches also help with text analytics and can be trained for entity extraction and other applications. This more modern form of natural language processing parses documents “as an aid to metadata classification, identity-related assets/concepts, measuring sentiment, and more,” Wills said.

Knowledge-sharing tools are also viable for other aspects of natural language technologies, including speech recognition. For example, if organizations had a taxonomy of all their proprietary products, services, and parts, they could “use taxonomies to train speech recognizers,” Aasman said. “Speech recognizers are really good at general English, but not good at product names.” This application opens up the realm of spoken conversations—meetings, customer phone calls, training sessions, etc.—to knowledge-sharing platforms. Notable offerings in this space involve natural language generation to summarize news, articles, and documents. These capabilities involve “deep neural networks for summary generations,” Chevala said. “This is the classic ‘so what’ kind of thing. You don’t want the end users to read the updated reports; we’re actually giving them the gist of that.”

Data governance

Amassing enterprise knowledge and sharing it between users would be much more risky than valuable without the proper data governance functions to ensure “whatever you find in the platform is trusted, reliable, and a source of truth,” Greenberger said. Data cataloging and some of the metadata management methods previously mentioned can pinpoint sensitive data and safeguard it with masking techniques according to governance policies. Some platforms also contain measures for data quality and data governance policy adherence—with analytics and reporting capabilities—so organizations can determine how their various knowledge management areas correlate to governance concerns.

Consequently, “The customer’s governance and knowledge management roles and processes determine how these capabilities are used by whom,” Wills said. A particularly effective means of managing the numerous regulations many organizations contend with is to create a knowledge graph linked to relevant datasets. This approach delivers “a declarative description in your knowledge graph that specifies, based on the types of data in your knowledge graph, what you can see and can’t see,” Aasman said. Preserving knowledge Enterprise knowledge continues to expand. Knowledge-sharing tools ensure this knowledge is preserved, curated, and easily disseminated throughout organizations. Their prioritization of collaboration, engagement, and interactions makes the growing decentralization and compartmentalization of the enterprise much more manageable.