Keep up with all of the essential KM news with a FREE subscription to KMWorld magazine. Find out more and subscribe today!

Enterprise search— an evolving technology

Article Featured Image

“To accommodate this type of organization,” said Jeff Evernham, vice president of product strategy for Sinequa, “we optimized the platform for Microsoft Azure. In addition, we have an integration with Microsoft Teams that allows users to provision applications within Teams and access information from anywhere. Our goal is to provide an open paradigm for collaboration applications and put search into the flow of users’ normal work.”

Sinequa has rebuilt its presentation layer in Angular, which is a modular, customizable interface that has become the industry de facto UI standard, according to Evernham. “We wanted to help expedite the development process by using open source components,” he noted. Developers can use this interface on top of Sinequa’s starter apps, which are packaged accelerators that do not require custom code. In October 2021, Sinequa announced three starter apps: the Search Starter App, which provides an out-of-the-box enterprise search application; the Sinequa Analytics Starter App, which extracts information from content; and the Usage Dashboard Starter App, which tracks usage of Sinequa applications.

A new capability in the works is neural search using deep learning. “What neural search can offer is increased relevance, especially where there is ambiguity,” Evernham said. “Its language model provides an understanding of the meaning in text that has not been possible before. Sinequa used a model similar to the one Google uses, but it is scaled back in order to require fewer resources. This technology is scheduled to be incorporated into the Sinequa Intelligent Search platform in early 2022.

Relevant self-service

The increased demand for self-service has been a significant driver for search. This demand was fueled in part by the shift to working from home, where individuals had to solve their own computer problems rather than being able to rely on corporate IT. A multinational computer networking company, for example, saw an increase in business as more people set up home offices. In light of pressure on its tech support staff, this company wanted to deploy a more extensive self-service capability. The company selected SearchBlox as its search engine to support this function and other search-enabled features.

SearchBlox is built on Elasticsearch/ OpenSearch, an open source search technology. Connectors to 300 data sources are built into the SearchBlox platform to allow easy access to information throughout the enterprise, and the search functionality can be embedded in any application. “Our design lets users do everything from one search box,” said Robert Selvaraj, CEO of SearchBlox. Among its features is SmartSuggest, which presents a list of options for the user once the user starts typing. The options presented can be based on frequent queries, promotional items, or personalization when the user clicks the search bar. “Our client’s customers can access information from FAQs, the site, or community sites,” continued Selvaraj.

The analytics in SearchBlox allows administrators to tune the results for relevancy. “Our client’s administrators can optimize the presentation of search results,” Selvaraj noted. If a lot of people are searching on a particular topic, the administrators can push a particular search result to the first page to expedite the search process. “We provide relevance tuning templates that can be applied differently to different parts of a website,” he added.

Another customer, a global reference laboratory, provides specialty lab testing in areas such as cardiology, endocrinology, genetics, and microbiology. Users can search for a certain test by name, but because the test names may be complicated, the SearchBlox application lets users just pick the first letter of the test name from an alphabet under the search box. A list of available tests is presented, from which the user can select the desired test. “If the user just knows the first few letters, they can easily find the test,” Selvaraj said. “We want each search session to be the only one that is needed, and for the user to find the answer on the first page.”

SearchBlox uses AI to handle repetitive tasks such as updating tagging and generating titles and descriptions for documents that do not already have them. “We provide a standard interface for chatbots, and guidance for building up a knowledgebase to connect the front end to the knowledgebase,” commented Selvaraj. “Like self-service, chatbots provide one more channel for delivering relevant information to users. However, relevant information can only be found if the underlying data is managed well.”

KMWorld Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues