Tips and Techniques to Close Knowledge Gaps

The pace of change in today’s world makes the emergence of knowledge gaps almost inevitable. They stem from technological change, introduction of new products, acquisitions and mergers that alter corporate structure, and scientific research. Staff changes through turnover and retirement also contribute. The challenge is to detect them quickly and fix them efficiently. Why worry about knowledge gaps? According to Panopto, companies with 1,000 employees can lose $2.4 million annually due to knowledge gaps, while companies with 30,000 employees can lose $72 million (rev.com/ blog/knowledge-loss).

One sure way to spot gaps is to monitor what types of information people are seeking but not finding. “Any good search engine has analytics to flag searches that produce no results or no click-throughs,” said Sid Probstein, founder and CEO of SWIRL. “What’s innovative is the ability to seek out new information to provide a response.”

Enterprise data contained within various applications is not always designed with findability in mind. SWIRL combines metasearch capabilities across enterprise data with AI and retrieval-augmented generation (RAG) to rapidly find the most relevant information, no matter where it is located or what format the data is in.

SWIRL adds AI-driven insights from the data to produce new content and analyses. In addition, SWIRL’s intelligent assistant provides tailored, individualized support, allowing users to automate routine tasks and run analytics. “SWIRL uses information about the user and their role,” Probstein added, “which helps users write a better question and get the best answer.”

SWIRL’s expansive reach, combined with natural language processing (NLP)  and generative AI (GenAI), enables employees to ask questions using conversational speech and receive accurate, relevant answers. “If someone in a company wants to know whether they are authorized to sign an NDA with another company but cannot find someone to answer the question, for example,” Probstein explained, “they can use SWIRL to answer the question.”

If the AI has been trained on, or has access to, legal standards and practices, it may tell the employee that they cannot sign the NDA, and then provide a link to send to the right person for further information. “The AI may know that the company has already signed an NDA,” Probstein commented, “and provide that information to the user. SWIRL can also generate follow-up questions and can track whether they were pursued by the user, to support additional insight into whether the issue has been closed.”

Within organizations, people spend a vast amount of time trying to answer questions but still face the problem of outdated, duplicative, and erroneous information. “AI offers the greatest potential for detecting and closing knowledge gaps, as well as verifying the accuracy of enterprise information,” concluded Probstein, “but we are only at the beginning of the AI journey. Before long, using AI to augment human intelligence and close the gaps in enterprise knowledge will be as natural as using a mobile phone is today.”

Navigating Software Processes

Knowledge gaps are often thought of in terms of content, but they can also occur in users’ understanding of processes. One example is learning the steps that need to be taken in using a software program, which can be a frustrating experience, resulting in an expensive enterprise investment.

The SouthEast Alaska Regional Health Consortium (SEARHC) was rolling out Workday, an ERP software product, to 1,400 users scattered across nearly 30 locations, most of which are only accessible by boat or seaplane. Aside from being logistically difficult, in-person training was considered less desirable because of the need for employees to be away from their jobs, absorb a great deal of information in a short span of time, and learn software out of the context of their work.