KMWorld 2024 Is Nov. 18-21 in Washington, DC. Register now for $100 off!

Enterprise search— an evolving technology

Article Featured Image

Enterprise search technology is an essential component of a knowledge-driven enterprise, allowing users to find the information they need and providing content to automated systems. It is an enabler for applications throughout the enterprise, including self-service, virtual assistants, and chatbots, all of which rely on retrieved information. Having an effective system for finding information that is scattered across many repositories has become even more important as the COVID-19 pandemic caused many employees to work remotely.

Enterprise search is a subset of search technology, which includes internet search engines, specialized products for ecommerce search, and voice search. These sectors all have different sets of vendors and different growth rates. According to Research and Markets, the enterprise search market reached nearly $4 billion in 2020 and is expected to grow about 11% per year, hitting $8 billion by 2027.

The importance of findability

As director of knowledge management at Takeda Pharmaceutical Company Limited, Giovanni Piazza is responsible for providing the content that its 50,000 employees and 25,000 contractors need in order to perform their jobs. An innovator in the field, Takeda spends a significant amount of its $30 billion in annual revenue on R&D, with about $30 million per year invested in external sources of scientific literature to support its research activities. Recognizing the need to have a robust search capability for this function and other corporate search needs, the company selected the Sinequa Intelligent Search platform as its enterprise search engine.

The content used by researchers comes from many different sites. To help them obtain and view the information they need in an efficient way, the Sinequa platform presents the content in one interface; therefore, the users do not have to become familiar with interfaces for the multiple destination sites on which the content is provided. Sinequa uses natural language process (NLP) query intent detection and other AI capabilities to return highly relevant results. Piazza considers Sinequa to be a good match for Takeda’ s needs. “It has the right capabilities, and the company provides high-quality professional services,” he said.

Rather than viewing the platform as a search engine, Piazza prefers to consider it as “a findability platform.” “Search is a technique,” he commented, “while findability is an outcome. This is core to my mission. The ability to find information is an imperative in pharma—for example, helping a researcher know whether we have done similar work in the past, which eliminates waste and redundancy. It also allows them to quickly discover who the experts are on a certain topic.”

Given the substantial costs of the scientific content, analytics that can reveal which data sources are most useful is an important part of the value that Sinequa’s platform provides. “We have developed a set of analytics that tells us the cost of each touch by the user,” Piazza continued, “whether it is to a page view, a URL, or a document. This allows us to place value judgments on any of the sets of content we purchase and optimize our spend.”

In addition, the metrics allow Takeda to get a picture of the userbase—who is using the content and what the patterns of usage are. “This process helps us find out if people who might benefit from the content are not using it, and whether we should reach out,” he noted. However, Piazza is cautious about enabling too much “push” in the content. “We prefer a pull approach in general, in which the user is the active party. Just because a user has accessed content on a certain topic does not mean they want dozens of additional articles about it.”

The Sinequa Intelligent Search platform is used broadly throughout Takeda to find information beyond what is needed for R&D activities. “Organizations have their cycles,” Piazza pointed out, “and we surface some information at certain times, such as benefit information when employees can make changes to their policies. At other times of the year, other information becomes relevant, and we surface it then to make it more visible to our userbase.”

Presenting relevant information is the positive outcome of an overall findability strategy. “There is a lot that needs to be done behind the scenes to make content accessible and findable,” observed Piazza. “Designing a system is not a static event, but an ongoing process that requires continued adaptation and continuous improvement. We view it as a journey, and we have a pipeline of many things waiting to be developed as part of our findability strategy.”

Sinequa’ s solution combines linguistic and semantic analyses, machine learning, and deep learning. It connects to numerous enterprise applications, including content management systems, enterprise resource systems, and customer relationship management systems. Its solution has been geared toward large, global entities and government organizations such as NASA that have many facilities.

KMWorld Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues