Klaus Fleischmann studied translation and interpreting in Vienna, holds an M.A. in Conference Interpreting from Monterey, California, and an MAS in Technical Communication from the Danube University in Austria. He started his career in the translation industry in a start-up agency in Monterey, California. After moving back to Austria, he decided to switch perspective and founded Kaleidoscope GmbH as software and process consulting company for Global Content. In 2007 he became managing director of Austria´s leading LSP, Eurocom Translation Services GmbH. Over the years, his focus narrowed more and more on enterprise terminology management. Together with his team, he now develops and implements Kaleidoscope´s Quickterm, an enterprise-level terminology management platform. Klaus served on the Boards of Directors of GALA, ELIA, Termnet and LTI, teaches at different Austrian universities, and gives numerous presentations and workshops on terminology at international conferences.

Terminology: The Clever Cure for Hallucinations

AI is the answer to everything, or so the common (mis)perception goes. But what about all the LLM hallucinations? What about the fact that GenAI does not adhere to your corporate language and produces inconsistent or simply wrong output? One potential cure may come as a surprise: your very own terminology! With terminology-augmented generation (TAG), AI can leverage the structured, human-grade, and multilingual data which is (often) available in a terminology database. TAG saves costs, improves precision, and reduces labor-intensive reworking of AI output. It is a game changer for techdoc, content, and localization teams, because it makes their terminology work essential for corporate AI projects and increases the value for the entire company. If you want to learn how to reposition your terminology as part of your company’s AI strategy and cure AI shortcomings, do not miss this presentation!

In this session, attendees will learn how to:

• Save time and money by making LLM output more correct, consistent and in line with the brand voice
• Prep the terminology data for use in LLMs or AI apps using terminology-augmented generation (TAG)
• Easily connect the termbase to any LLM or AI tool using the Model Context Procotol (MCP)
• Get a seat at the corporate AI strategy table by positioning themselves and their terminology data
• Stay valuable as information architects in the age of AI
• Present the case for terminology augmented generation to the CIO or whoever is in charge of the AI strategy