5 Ways Text Analytics and NLP Make Internal Search Better

Content Intelligence

Content intelligence provides employees with critical knowledge locked within documents, including the reason they exist and their most important attributes. Imagine a dashboard that provides:

♦ Information about the file itself (who created it and when, who last edited it and when, number of versions, etc.)

♦ The textual content of the document (key concepts, document summary, named entities, etc.)

♦ Multimedia information (images and videos, including identified text and transcripts, objects, etc.)

♦ Taxonomy information (company categories, specific terms, data classification, etc.)

Imagine a search experience that provides users with access to important information about the documents that show up in their search results. Taking advantage of AI technologies such as natural language processing, image/video processing, extracted keywords and phrase search can deliver important intelligence about documents, without the need to open and read them. Although this can be done individually with each system, a simpler approach is to implement it as part of enterprise search with a single index and single UI.

Intelligent Search Bot

Search bots are all over the internet, but not many are used for internal search. A search bot combines a digital assistant-type user interface with natural language processing, a range of text analytics capabilities, and machine learning to act as a search assistant or even replace the search bar so users can find relevant information faster than using keyword search.

Below are scenarios in which a search ?bot can help users quickly find relevant ?information:

♦ Too many results:

  • The bot recommends alternate queries or refiners to narrow down the result set, even showing the number of results per option.

♦ Too few results:

  • The bot recommends alternate queries or the removal of refiners to deliver better results.

♦ Conversations:

  • The bot recognizes the question the user is asking in natural language and converses to fully understand what is desired to provide a precise result. For example, a user may be looking for support cases and the bot can ask the desired priority level to narrow results.

Bots automate multiple search processes, providing assistance for even the most novice users so they can find the right information quickly.

Sentiment Analysis

Sentiment analysis is used heavily in customer-facing applications. However, the same doesn’t apply to employee-facing applications because the majority of information inside organizations is captured in documents, for which sentiment is rarely relevant. Content types for which sentiment analysis can be meaningful include:

♦ Social technologies like Yammer, Teams and Workplace by Facebook.

♦ Emails and other electronic communications.

Your employees are talking, using collaboration platforms such as Yammer and Workplace by Facebook, but how productive are those conversations? Sentiment analysis helps you improve employee engagement and morale by identifying, and taking action to correct, areas of employee dissatisfaction.

KMWorld Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues