Navigating the intersection of KM and GenAI: Challenges and solutions

Implementing generative AI (GenAI) at the enterprise level is—while desired by many executives—a difficult, infrastructure-defining task. Despite its promises, achieving tangible success from GenAI requires a deft hand (or, more aptly, a deft knowledge system) to begin delivering value.

David Seuss, CEO of Northern Light, joined KMWorld’s webinar, Overcoming the Barriers: How to Integrate Generative AI into Knowledge Management Systems, to explore the nuances of leveraging GenAI to enhance KM systems and overcome the various challenges associated with its deployment.

Seuss opened by explaining that the pressure to adopt GenAI is coming from above, where, according to a 2024 Accenture report, 95% of executives believe GenAI will compel their organization to modernize its architecture. Furthermore, according to Gartner, one of the largest applications for GenAI is within the realm of research, including the discovery, curation, and summarization of information.

With other reports detailing the utility of GenAI for business strategy—namely, Harvard Business School and BCG’s collaboration which found that groups using GenAI finished tasks 25% faster with 40% higher quality—its potential value is quite clear.

Competitive intelligence is Northern Light’s focus, offering SinglePoint, the custom-built enterprise knowledge management platform that seamlessly integrates and enables full-text search of all enterprise research resources. With technology like SinglePoint, the benefits of GenAI for market research are numerous, including:

• Much better business research outcomes due to insights from more document and reallocation of time
• Better use of high-value content through reduced effort to access content insights
• Faster time to insight
• Improved marketing, product development, and corporate strategy decisions

Yet, despite this range of possibilities, there are four main barriers to success, noted Seuss: content, accuracy, governance, and security.

In the realm of content, it is important to remember that competitive intelligence content comes mostly from the outside world. Content preparation becomes a priceless asset, helping to make research content as useful as it can possibly be. Enterprises should curate relevant content by:

• Extracting text and capturing metadata
• Indexing for search
• Creating GenAI text database and process flow to deliver text to the large language model (LLM) at run time

Seuss further cautioned webinar viewers in the nature of web content, an unruly, messy amalgamation of information that, while insightful and timely, is a breeding ground for copyright infringement risks. To combat this risk, Seuss emphasized the need to aggressively curate content so that it cites and links back to its source material. That way, not only can the user vet the source content for themselves, it helps drive copyright compliance.

While content preparation for GenAI takes many skills to accomplish, it is vital for cultivating a robust knowledge foundation from which GenAI can succeed, according to Seuss.

Seuss then explored various other GenAI challenges and solutions, including using retrieval augmented generation (RAG) for mitigating against hallucinations, governance processes to assist in GenAI maintenance, and more.

For the full, in-depth webinar, you can view an archived version of the webinar here.