Transforming Knowledge Management in Businesses with Generative AI: Opportunities and Challenges

In today’s rapidly evolving business landscape, the need for efficient knowledge management (KM) has never been more critical. Companies are overwhelmed with vast amounts of data and information that must be effectively managed and utilized to maintain a competitive edge. Traditional KM methods often fall short, leading to inefficiencies and missed opportunities. Generative AI is emerging as a transformative technology that is revolutionizing the way businesses handle their internal information, driving productivity and fostering innovation.

Generative AI, particularly through the use of Large Language Models (LLMs), enables companies to enhance their KM systems by automating the capture, retrieval, and generation of information (Khan et al, 2024). This not only saves time and resources but also provides employees with the tools they need to make informed decisions quickly and accurately. There are several advantages of generative AI in KM, including improved knowledge workers’ productivity, enhanced decision-making capabilities, and the ability to leverage organizational knowledge to drive strategic initiatives.

To further explore the transformative potential of generative AI, a hybrid seminar titled “Generative AI-Enhanced Knowledge Management in Business” was organized at Haaga-Helia University of Applied Sciences, Helsinki on 8th May 2024. This event, funded by the European Regional Development Fund (ERDF) and hosted in collaboration with the Ulysseus Applied AI for Business and Education Innovation Hub, brought together experts from various sectors to discuss the integration of AI into business KM. The seminar was designed for medium-sized companies looking to leverage their information resources, AI solution providers exploring new opportunities, and AI experts interested in sharing their knowledge.

The seminar comprised presentations as well as a panel discussion. The presentations focused on the opportunities offered by generative AI for efficient KM in business. The panel discussion further explored the applications, integration, benefits, and risks of generative AI in business, along with strategies to address these challenges. The panel discussion was moderated by Dr. Umar Ali Khan, a senior researcher at Haaga-Helia University of Applied Sciences. The panelists included Alexander Finn, Head of Technology at SiloGen; Elina Seppälä, Head of Knowledge Management at RELEX Solutions; Severi Tikkala, CTO of Confidential Mind, and Johanna Mäkeläinen, a senior lecturer in marketing and AI Business trainer at Haaga-Helia University.

We will share the insights from the panel discussion through a two-part article series. The first part will discuss the opportunities, use cases, and transition challenges of integrating generative AI into business knowledge management. In the second part, we will explore alternative technological solutions, the associated risks and regulatory challenges, and the critical success factors for effective implementation of generative AI systems in businesses.

Opportunities of Generative AI for Business

Most organizations, particularly larger ones, accumulate vast amounts of knowledge over time – they collect and store a myriad of documents, build internal social networks, and maintain internal wikis. But all this knowledge is difficult to access. According to recent analytical reports. knowledge workers spend an average of 2.8 hours a week looking for or requesting information and 2.0 hours a week recreating information and work that already exists elsewhere in the organization (Trees, L., 2022). Generative AI holds immense potential for addressing this issue.

The impact of generative AI in business can be identified on three distinct levels: individual, team, and organizational. At the individual level, knowledge workers experience increased productivity with generative AI. Whether it’s searching for information, planning tasks, or any other online activities, AI acts as a helpful assistant, streamlining daily operations and enhancing efficiency. On the team level, generative AI amplifies the collaborative efforts of experts working together. For example, in marketing, AI can assist in sketching new images and drafting initial versions of materials to speed up the planning process and improve the overall team’s output. At the organizational level, which is the most complex, the strategic integration of generative AI is still in its early stages. Many companies are beginning to recognize the importance of developing a comprehensive AI strategy. An increasing number of businesses seeking training sessions are eager to learn how to incorporate AI into their operations strategically. This growing interest highlights the shift towards integrating AI at a strategic level within organizations.

Generative AI Use Cases for Knowledge Management

As generative AI becomes integral to business, companies should consider how to leverage it for efficient KM as a competitive edge.

One of the most popular use cases of integrating generative AI into business is the “talk to your company knowledge” solution. This involves integrating multiple knowledge sources into a search index, which is then coupled with an LLM to create a chatbot. This chatbot can efficiently answer queries by accessing and synthesizing information from various internal databases. The most impactful and widely used example of this solution is in the customer support function. The companies that provide support services to their customers face the challenge of conducting extensive and time-consuming training for their customer support agents. Despite having scripts, agents need to understand the context to deviate appropriately, especially when supporting customers from multiple different companies (in the case of B2B), requiring quick context switching. Generative AI provides a transformative solution by enabling agents to quickly retrieve accurate answers to user questions. This capability significantly reduces the time and effort spent on continuous retraining, allowing agents to rely on the existing knowledge base.

There are even more sophisticated use cases of generative AI. For example, an online marketing platform might use an LLM to transform a natural language description of a marketing campaign into a fully configured campaign. This process can include generating the campaign’s text, images, and email configurations. By specifying target audiences, such as digital native companies in the Nordic region, and desired timing, like weekdays, the LLM can automate the creation of a tailored marketing campaign.

Companies can use GenAI to automatically generate analytical reports from internal data, providing valuable insights with minimal manual effort. Another internal data-driven use case of GenAI is performing knowledge gap analysis, where the system identifies areas where information is lacking or outdated. These applications offer significant potential for enhancing business processes and decision-making. The ability to automatically synthesize and analyze large volumes of data can transform how businesses manage knowledge and maintain a competitive edge in their respective industries.

Information or fact extraction is a very promising area of focus. LLMs can be used to automate the extraction of specific details from vast amounts of data. For instance, a company monitoring updates on road infrastructure can use LLMs to scan news articles and social media posts for relevant information, such as the closure of a bridge or a traffic accident. This automation reduces the manual effort required and ensures timely and accurate data extraction.

A successful example of creating a generative AI assistant for accessing organizational knowledge is demonstrated by Rebot tool of RELEX Solutions (RELEX Solutions, 2024), a provider of unified supply chain and retail planning software solutions. The tool has streamlined access to information, making it easier and faster for employees, customers, and partners to find what they need. An internal survey revealed a 19% increase in productivity in 2023, with estimates indicating a further 32% gain for 2024-2025 (RELEX Solutions, 2024). While the company has not yet quantified the impact on customer experience, they anticipate similar improvements as customers begin to find information more quickly and efficiently. Though this solution has proven effective, it does come with some limitations. For instance, when dealing with very similar documents from different perspectives, users must exercise careful prompt engineering to ensure they retrieve the correct and relevant information.

RELEX’s customers primarily use Rebot to understand RELEX’s highly configurable software products. They seek to maximize the benefits and explore the various possibilities that the software solutions can offer for their businesses. Typically, their customers are interested in searching for detailed information or reference materials that help them comprehend the functionality and capabilities of the software. RELEX’s customers often inquire about how specific features work, what is involved in the backend processes, and how they can leverage these features to optimize their operations. Rebot provides this information in a flexible and accessible manner, enabling customers to gain deeper insights and make more informed decisions about using RELEX’s software solutions effectively.

Transition to Generative AI Systems and Change Management

Companies face various challenges when transitioning to generative AI-based systems for KM. One of the main challenges is change management and integrating the new AI system with existing systems. This process can be complex and requires a strategic overhaul of current content management practices. The amount of content generation sources and channels is growing, and there is a need to deliver relevant content to various users at the right time and place. Companies might need to rebuild their content management systems to provide consistent and relevant content for various audiences. This rebuild is necessary to ensure that the content is up-to-date and valuable for different user groups.

Technology providers such as Confidential Mind are trying to address these challenges of user companies. They are creating software products for system integration and technological change management, ensuring that data used in generative AI systems remains updated and properly indexed. This involves setting up databases and implementing triggers to monitor changes in data sets. Flexibility is crucial, allowing the addition or removal of data as needed. User companies can access the necessary data through a single API, enabling them to focus on leveraging the data rather than managing it. This approach not only facilitates smooth transitions to AI-based systems but also ensures that these systems remain adaptable to evolving business needs.

Another challenge is maintaining the credibility and accuracy of the information provided by AI systems. Users need to trust the responses generated by chatbots or other AI tools. Links to the original source documents must always be provided to address this. This transparency allows users to verify the information and understand its context by checking the original content. Ensuring that source content is readily available and easily accessible is crucial for maintaining trust and reliability.

An incremental and thoughtful approach must be applied to change management. Starting small with non-critical projects allows organizations to gradually adapt to the new technology and understand its internal implications. The first step is to identify and start with simple, clear use cases where AI can have a significant impact. Businesses should not be afraid to fail; instead, they should adopt a mindset of continuous learning and improvement. By iterating on their initial implementations, businesses can gradually refine their AI applications. In addition, businesses can also face challenges in establishing governance and oversight for AI tool usage. The initial challenge could be determining who can authorize the use of AI tools and how these tools should be managed. This involves regulating who uses the tools.

In the case of Rebot, RELEX began with small, manageable sets of information, primarily focusing on customer-facing content such as online help resources. This approach allowed them to study and recognize what information their customers needed and what gaps existed. For their internal Rebot, they gathered existing information within the company, ensuring it was up-to-date and valuable for all employees. To integrate this knowledge into an LLM, they started small and expanded as they learned more about their customers’ needs and the most effective ways to meet them. This incremental approach allowed them to build a robust and useful AI tool that could grow and evolve with their business requirements.

Despite these challenges, it’s crucial not to hinder employees from benefiting from AI tools at an individual level, given their potential to significantly increase productivity. Therefore, the goal is to manage the integration methodically, ensuring clear guidelines and regulations, while also enabling employees to leverage AI to enhance their work. This balanced approach helps in maximizing the productivity gains from AI while maintaining control and compliance within the organization.

Conclusion

Generative AI offers transformative potential for business knowledge management and provides innovative solutions to long-standing challenges such as knowledge capture, synthesis, access, and decision-making support. The opportunities for enhancing productivity at the individual, team, and organizational levels are substantial. However, realizing these benefits requires careful consideration of the specific use cases and a strategic approach to transitioning to generative AI-driven systems. Change management and the integration with existing systems are important to ensure that the adoption of generative AI is both effective and sustainable. In the next part of this series, we will discuss the alternative technological solutions, the risks and regulatory considerations, and the key factors that can drive successful generative AI integration for business knowledge management.