Text Analytics Forum is part of KMWorld 2024 featuring five co-located events: KMWorld 2024, Taxonomy Boot Camp, Enterprise Search & Discovery, Text Analytics Forum, and Enterprise AI World. Upgrade to a Platinum Pass for your choice of two preconference workshops or access to Taxonomy Boot Camp on Monday. Workshops are also separately priced.

Join us for the Enterprise Solutions Showcase Grand Opening reception. Explore the latest products and services from the top companies in the marketplace while enjoying drinks and light bites. Open to all conference attendees, speakers, and sponsors.

Text Analytics Forum is part of KMWorld 2024. Upgrade to a Platinum or Gold Pass for extended access to KMWorld2024, Enterprise Search & Discovery, Taxonomy Boot Camp, and Enterprise AI World, a series of co-located events happening alongside Text Analytics Forum 2024. See the registration page for details.

Stop by the Enterprise Solutions Showcase after a full day of stimulating sessions to mix and mingle with other conference attendees, speakers, and our conference sponsors.

Today’s business landscape is changing faster than ever before in history. The power of inclusive engagement and collaborative curiosity cannot be overstated. Join leadership strategist and award-winning author Dan Pontefract as he unveils essential techniques to cultivate a culture where every voice is valued and heard. Drawing on global primary research and more than 25 years of experience with leading organizations, he shares actionable insights and transformative strategies that empower KM teams to work better together. Gain expert tips on creating an environment where collaboration prospers, every team member feels genuinely valued, and the opportunity to bloom is open to all.

As organizations integrate AI into their product ecosystems, innovative KM practices are essential to keep information relevant and useful. In the age of GenAI and LLMs, the principle of “garbage in, garbage out” remains true—AI systems are only as effective as the data they process. Chmaj discusses emerging content models, new competencies, advanced authoring techniques, and governance practices that are transforming the KM landscape. It’s vital that companies evolve their technology, resources, and strategies to unlock the full potential of AI-driven KM capabilities. In this dynamic and complex AI landscape, content remains king! Get lots of insights and ideas from our experienced KM leader.

The productivity of knowledge workers is critically important to the growth and profitability of businesses. However, they remain weighted down, spending nearly half their time on mundane tasks, leaving less time for the work that matters most. An organization’s information is the lifeline that provides the insights required to gain collective intelligence. Speakers discuss how organizations can gain a strategic, competitive advantage by leveraging knowledge work automation. They share actionable insights on harnessing the power of automation and AI to eliminate information chaos, improve productivity, and reduce business risk to enable knowledge workers to thrive. 

Generative AI and large language models (LLMs) are reshaping the landscape of corporate knowledge management, heralding a new era where technology speaks our language and enhances our work. Over the past year, retrieval-augmented generation (RAG) has emerged as the premier method for applying LLMs to corporate content, proving its value in real-world applications. However, the journey doesn't end with RAG; in fact, it’s just the beginning. Evernham shows how AI is transforming KM with real-world examples of organizations' use of RAG to unlock the full potential of their knowledge; introduce our new digital colleagues, AI-powered assistants that work alongside knowledge workers, to enhance productivity and decision making; and peeks into the future with a look at multimodal AI and agents poised to perform complex tasks and accelerate innovation.

What are the current and future trends for the field of text analytics? Join program chair Tom Reamy for an overview of the conference themes and highlights and a look at what is driving the field forward. The theme this year is New Synergies, New Solutions. As the hype around GenAI subsides somewhat, people are asking about how to build solutions that deliver real business value. Join us as we explore how text analytics is an essential tool for adding accuracy and depth to GenAI and opening up new possibilities. We are also bringing back our popular Ask the Experts panel.

This presentation explores the transformative potential of GenAI in business decision-making. It highlights the shift from traditional, chat-based interfaces to more autonomous, AI-driven systems, demonstrating how businesses can leverage this technology to enhance efficiency, accuracy, and strategic insight.

Which algorithm is right for you? The one that gets the job done. Technologies are only as good as their results, and this requires fitting the right approach for your problem. Downs demonstrates the ways in which you can assemble a diversity of human-curated and machine learning tools at your fingertips, selecting the right tool for your task. Use cases explored include assembling enterprise taxonomies, NLP-driven text analytics, and LLMs to create associations between concepts and content, including novel entity extraction; marrying enterprise ontologies and knowledge graphs with LLMs to power retrieval augmented generation (RAG) to support business process automation; and storing outputs of machine tagging on content-aware knowledge graphs for further functionality and insight.

As GenAI continues its rapid evolution, the extent of its impact and the number of industries disrupted are increasing at an equally rapid pace. Whether the disruption represents a crisis or opportunity is up to us. But first, we need to step away from the hype, take a look under the hood, and see what’s really going on inside. Only then can we rationally think about how it might fit into the big picture. That picture includes not only generative and extractive AI, but also text and data analytics, graph databases and knowledge graphs, layered ontology, and the all-important but often-neglected incorporation of human intuition, sense-making and expertise. By carefully stitching all of these components together, major leaps in performance can be achieved. This session presents a pilot configuration along with the data and knowledge flows among these various components, test results showing problems encountered and how AI can learn from humans and vice versa, and plans for future enhancements.

Text analytics has been around a lot longer than GenAI, and the technologies serve different purposes. But the combination of the two offers exciting potential for businesses. For instance, consider the possibilities of marrying the trend analysis and insight extraction capabilities of text analytics to GenAI's text summarization capabilities. The result might be thought of as a "consultant in a box"—a powerful fusion that promises to dramatically enhance productivity in functions like strategic analysis and planning. In this session, Seuss paints a picture of a future in which GenAI reports on the insights that text analytics generates.

With the revolution and continuous introduction of AI technologies (e.g., ChatGPT, LLM), how should one consider an effective integration and responsible use in day-to-day effort for text analysis? What are some of the synergies and watchouts that one should consider to deliver greater value and efficiencies for effective decision making. This session talks about how best to adopt the hybrid text analytics (the intelligent interplay of human and software) and mindset to achieve the optimal outcome in a responsible way.

This presentation focuses on the application of LLM-powered GenAI in improving knowledge discovery within our open access publications catalog. The integration of this technology facilitates more intuitive and effective interactions with our extensive corpus of over 13,000 publications. The LLM plays a role similar to that of a generalist, providing broad overviews and insights across a wide array of topics. However, achieving precision and accuracy necessitates the detailed contextualization that metadata and knowledge graphs provide. These elements provide context akin to that of specialists, ensuring each piece of information is accurately contextualized and enhancing the overall reliability of the discovery process. We use a retrieval-augmented generation (RAG) approach, leveraging advanced metadata and knowledge graphs to enrich AI-generated responses with relevant context to make sure users receive appropriate insights that reflect the nuances and complexities inherent in our publication corpus. We highlight specific examples of how this approach underscores the complementary roles of the LLM and metadata, exemplified in the specific case of our publications catalog. The LLM’s broad capabilities work together with the detailed, contextual expertise provided by metadata and knowledge graphs to deliver a more precise, accurate, and “truthful” discovery experience.

Government organizations are responsible for reviewing public commentary to proposed regulations and responding to all comments, which can take thousands of hours. This presentation offers a solution to address this challenge by automating the process with text analytics and GenAI. It highlights how text analytics serves as a method to calibrate the data which is fed into GenAI models to enhance accuracy and time to value and provides a layer of traceability to summaries. It also highlights how the process can apply to any number of commercial and public sector use cases.

GenAI is poised to revolutionize industries, but are you ready to harness its potential? This session goes beyond the hype, equipping you with a practical framework to assess your organization's GenAI readiness. In this interactive session, attendees discover why a thorough readiness assessment is the secret weapon of successful GenAI adopters; gain crucial insights into the data management, AI capabilities, and technological infrastructure needed for a smooth transition; unveil a structured framework to evaluate their organization’s preparedness across these key metrics; identify strengths to build on and weaknesses to address before GenAI implementation; turn challenges into opportunities; and learn how to bridge identified gaps and set your organization on a path to leverage GenAI for a sustainable competitive advantage. Attendees walk away with the tools and knowledge to confidently assess their GenAI readiness and propel their organization toward a data-driven, AI-powered future.

A panel of four text analytics experts answer questions that have been gathered before the conference, during the conference, and some additional questions from the program chair. This has always been one of our most popular features, so come prepared with your favorite questions and be ready to learn.

In today’s landscape, the airwaves resonate with discussions about AI. Embracing emerging technologies has elevated KM efforts. From the early days of portals and expertise location to the collaborative power of wikis and shared spaces, technology has consistently given KM a much-needed boost. This conversation begins with Hubert discussing how experts fit into the AI equation with knowledge creation, capture, and seamless transfer. Given observations from today’s breakneck speed of business, the challenges have only gotten more daunting and the need for tapping into expertise is needed even more. She shares a model for thinking of how to enable AI using the skills of experts. In the dynamic interplay between human expertise and AI, she provides a path to unlock the potential for groundbreaking insights that lead us into uncharted territories of new knowledge. As we navigate this synergistic landscape, both seasoned experts and the KM programs that bolster their efforts must seize the opportunity to capitalize on their collective wisdom. Those who work smarter, fearlessly embracing collaboration with AI, are poised for success. 

Is your enterprise experimenting with generative answering, but facing significant challenges with data cleanliness? Olguin shares best practices and discusses the strategic role KM plays in delivering effective GenAI. Get proven strategies for refining generative outputs and practical insights from enterprise customers including Xero, F5, and Forcepoint. Uncover proven metrics and KPIs to ensure accurate, relevant, and safe generated answers, optimizing your knowledgebases. Enhance productivity, proficiency, and decisionmaking with curated generative answering.

Coleman cuts through the chaos and offers a practical road map for KM teams grappling with the AI revolution. He discusses three key steps to make no-regrets AI investments while safeguarding your hard-earned KM progress: how to separate AI fact from fiction; how to identify AI opportunities that actually complement your existing KM ecosystem; and how to implement a measured, value-driven approach to AI adoption. Get actionable insights to navigate the next phase of KM with confidence.

We have all been wowed by amazing examples of generative AI. But beyond the headlines, things look a little different. While employees are leveraging generative AI to increase productivity on individual and team levels, enterprises are frequently not yet using generative AI in a structured manner across their organizations, mainly due to the lack of private enterprise data. Capturing tacit expert knowledge is where big quantifiable gains can be made. The founders of Sugarwork share a customer case study to highlight the productivity gains the business experienced when they captured and deployed tacit expert knowledge at scale using generative AI.

For all the success that GenAI has accomplished in the public sphere, it has a number of limitations within the enterprise that text analytics can overcome—if done correctly. This session describes how text analytics’ auto-categorization and data extraction can overcome the five well-known limitations of GenAI and, at the same time, shows how the combination can greatly enhance a wide range of enterprise applications. For example, while RAG can help with hallucinations, the search component is only as good as the basic enterprise search, which we’ve seen needs text analytics to really shine. We cover how to develop a multidimensional platform that can be used to build multiple applications that reflect your world, not a generic and overly simplistic public world, fact-check hallucinations, provide human-understandable explanations, check security, and create a highly accurate training set for enterprise LLMs.

LLMs like GPT-3 or Llama have demonstrated remarkable capabilities in generating text on a wide range of topics. However, these models are not without limitations. A major challenge is ensuring that the information generated is accurate and up-to-date, especially for rapidly changing or highly specialized domains. While “slightly inaccurate” information can be OK in some areas, it can be disastrous in medicine, law, engineering, and other fields in which accuracy is mission-critical; these are areas in which the nightmare scenario of LLMs hallucinating and providing false information isn’t only annoying, it can be deadly. Retrieval-augmented generation (RAG) is a technique that allows LLMs to incorporate external information from a corpus of documents during the text generation process. RAG mitigates the hallucination problem by grounding LLM responses in verified data sources. The key components of text analytics—text preprocessing, natural language processing, entity extraction, and structuring data—play a critical role in this process. These techniques ensure that data fed into an LLM is clean, relevant, and structured to maximize the accuracy and reliability of the responses. By leveraging organizational knowledge and domain-specific datasets, RAG significantly enhances the performance and trustworthiness of LLM outputs. This presentation examines the intricacies of data and content preparation within the RAG architecture to improve LLM performance. Gross walks through a data scenario, illustrating the step-by-step process of how data is preprocessed, structured, and integrated into the RAG framework. Attendees gain insights into best practices for data preparation, the benefits of integrating text analytics into AI workflows, and practical strategies for deploying RAG to enhance the accuracy and reliability of GenAI systems.

HealthStream maintains a dataset of regulatory requirements for professional licensing for healthcare professionals and provides courses to meet many of these requirements. It wants to explore a regulatory recommender that would use the tags applied to courses and regulations to pair an individual with the appropriate training material specific to their license, care setting, and state. This is being built in collaboration with Semantic Web Company, utilizing its LLM-based Taxonomy Advisor and Recommender feature to extend the taxonomy and its Semantic Integrator technologies to power the recommendations. The primary objective for this recommender is that its architecture be easily understandable and maintainable by our knowledge management team. A previous attempt to build such a recommender was too complex, leading to unexplainable results and an inability to fine-tune the recommendations based on feedback. It is critical that these recommendations be accurate, as this is legal content that has implications for professionals maintaining licensure if incorrect courses are suggested to fulfill the requirements.

Today, enterprises have more tools to create and share information than ever, leading to significant challenges in managing duplicate content. This session explores how a large supply chain organization implements green information management to support sustainability goals. Presenters from Enterprise Knowledge showcase a hybrid AI framework, combining heuristic and LLM-based approaches to effectively analyze and reduce duplicate content across enterprise repositories at scale. This approach identifies duplicates and near-duplicates and enhances content retrieval, ensuring information accuracy and reducing storage costs. By leveraging AI for content analysis, organizations can streamline their data management processes and contribute to their environmental, social, and governance (ESG) goals by minimizing the carbon footprint associated with redundant data. The presentation demonstrates the environmental benefits of reducing duplicate content, focusing on carbon footprint reduction. It also addresses how this information generates aggregate statistics and resultant carbon footprint, pushing for a cultural shift toward greener information management. Attendees gain insights into the practical implementation of AI-driven content analysis frameworks, the environmental impact of effective data management, and the importance of integrating ESG goals into information management strategies. This approach promotes a cultural shift toward sustainable data practices, significantly improving data management efficiency. Attendees learn how modern AI techniques can transform their enterprise’s data practices and support a sustainable future.

Classifying content based on thesauri is an established method in content strategy and information management. Knowledge representation, including formal logic and terminology control, provides precise and consistent indexing, enhancing information retrieval. The advent of LLMs has given these approaches a significant boost. Effective GenAI strategies, as suggested by Gartner and other analysts, now require efficient and accurate labeling of corpora at unprecedented volumes and speeds. Addressing the challenges of unlocking unknowns from latent content, guiding users in discovering new insights, and preventing hallucinations in GenAIs have become increasingly critical. As enterprises explore merging LLMs with symbolic AI, graph grounding—the formal representation of information and data—emerges as a pivotal challenge. This session introduces key concepts and presents interactive and automated tools for annotating, classifying, and indexing unstructured text and documents. A central concept is inference tagging, a method that extends document annotation beyond explicit concepts by deriving new information from ontologies and rules. Gabler showcases use cases demonstrating how content can drive risk identification, personalized recommendations, and decision support, highlighting the practical applications and benefits of these advanced techniques.

This case study outlines the development of an information discovery application that integrates a knowledge graph with an LLM. The presentation focuses on demonstrating how well-established knowledge management techniques can enhance the effectiveness and user-friendliness of contemporary retrieval-augmented generation (RAG) information retrieval systems. Additionally, it highlights how the application meets the stringent production requirement typical for regulated enterprises, ensuring relevance, compliance and security. It focuses on the innovative techniques to overcome traditional obstacles in information retrieval. These challenges arise primarily because users frequently submit (natural language) queries that are poorly constructed, vague, ambiguous, lack clear and adequate context, and/or contain narrow, domain-specific terminology. By addressing these issues, the applied approach is able to measurably improve the discovery of relevant information while at the same time increasing the ease of use. At its conclusion, an overview is shared of the measured business and user benefits from the increased findability, understanding, trust, and ease of use of the assets of the organization and how their increased reuse and repurpose capabilities affect the daily work of both knowledge workers and specialists.

As AI becomes more ubiquitous, enterprise leaders face a challenging dilemma. On the one hand, they need to leverage the power of AI to increase productivity and drive innovation. On the other hand, they must protect their organizations against the risks of open generative AI solutions. Dallhoff details the advancements and risks of generative AI in the workplace, including critical issues like data leakage, hallucination, and legal consequences. He offers insight into how enterprises can safely harness AI by adopting thoughtful approaches, like the ones his company prioritized when building the Pryon platform.

KMWorld magazine is proud to sponsor the 2024 KMWorld awards, KM Promise & KM Reality, which are designed to celebrate the success stories of knowledge management. Information Today, Inc. is excited to introduce and present a new award: KM Community Award recognizes an individual who has made a significant impact in the KM community. The awards will be presented along with Step Two’s Digital Awards, where you get a sneak peek behind the firewall of these organizations. Find out more—kmworld.com/Conference/2024/Awards.aspx

In the era of big data and AI, accurate and efficient terminology matching is critical for numerous applications, including information retrieval, data integration, and semantic search. This presentation introduces an innovative approach to intelligent terminology matching by leveraging a knowledge graph in conjunction with a vector database. Our method addresses the challenges of synonymy, polysemy, and contextual relevance in terminology matching. A knowledge graph provides a structured and semantically rich representation of entities and their interrelationships, offering a powerful foundation for understanding and disambiguating terms. By integrating a vector database, the system's ability to capture semantic nuances is enhanced through high-dimensional vector embeddings. These embeddings are generated using advanced natural language processing techniques, such as transformer models, to encode the semantic content of terms within a multidimensional space. This approach consists of three main components: 1) constructing and enriching the knowledge graph with domain-specific terminology and relationships, 2) generating vector embeddings for terms using pretrained language models, and 3) performing similarity computations to match terms based on their semantic proximity within the vector space. This combination enables precise and context-aware terminology matching, significantly improving the accuracy and relevance of results compared to traditional keyword-based methods. Aasman discusses the effectiveness of our method across various domains, highlighting best practices in handling complex and diverse terminologies. Attendees gain insights into the technical details, practical implementation, and potential applications of this cutting-edge approach to terminology matching.

Ananyan provides a step-by-step tutorial demonstrating the creation of a GenAI-based solution for VoC data analysis. Attendees learn the basics of prompt engineering, see how to adjust prompts to improve quality of results, what techniques help eliminate hallucinations, how to enhance the value of results with text analytics, and what additional engines can be used to produce an easy-to-comprehend summary of results for decision makers.

Embark on a journey through the highs and lows of developing GenAI for the enterprise in an era where technology moves at breakneck speed. Zeta Alpha dove headfirst into GenAI. By January 2023, it had successfully deployed features into production. Yet, this journey was not without its challenges and invaluable lessons. Join our speaker as he shares insights from his own development of Zeta Alpha's public Discovery platform in AI R&D, as well as from assisting enterprise clients across diverse sectors like chemical, HR, high-tech industrial, and regulatory industries on their path from prototype to production. Discover the pitfalls in treating GenAI projects as standard software development endeavors and learn how to avoid wasted effort on solutions that will soon be provided by LLM vendors. Explore the nuances of collaboration with business stakeholders to ensure that the GenAI features you develop are both cutting-edge and genuinely beneficial for their operations. Gain insights into expectation management, strategic technology selection, and the pivotal role of business alignment in project success. Don't miss this opportunity to learn from Zeta Alpha’s journey and equip yourself with strategies to navigate the complexities of putting GenAI into production, steering clear of potential pitfalls along the way.

This session tackles the crucial role of content preparation for AI in KM. Discover strategies for optimizing content for AI-powered features such as generative search. Hill delves into everything from data structuring to crafting effective prompts equipping you to unlock your knowledgebase's true potential. See AI in action with a compelling case study showcasing a real-world application of generative search and its significant impact. Learn how to transform knowledge boundaries into a springboard for limitless growth.

Where is KM going with all the AI developments for the enterprise? How are our organizations responding to the social structures and changes in our world? How are they innovating and exceeding customer expectations? Get inspiration from our practitioners and futurists and be ready for KM in 2025.