{"id":10798,"date":"2025-05-22T12:51:50","date_gmt":"2025-05-22T10:51:50","guid":{"rendered":"http:\/\/ic-steiermark.at.w02242c6.kasserver.com\/news\/review-this-was-the-ai-update-for-export\/"},"modified":"2026-11-03T15:40:04","modified_gmt":"2026-11-03T14:40:04","slug":"review-this-was-the-ai-update-for-export","status":"publish","type":"news","link":"https:\/\/www.ic-steiermark.at\/en\/news\/review-this-was-the-ai-update-for-export\/","title":{"rendered":"Review | This was the AI Update for Export"},"content":{"rendered":"
\n
\n \n
\n \n \"Logo\n <\/a>\n \n \"\"\n <\/a>\n <\/div>\n <\/div>\n <\/div>\n
\n
\n
\n
\n \n

“It is not about the AI tools, but about the business processes.” \u2013 this sentence from long-standing ICS cooperation partner and AI coach Lothar Lackner<\/strong> stuck, and rightly so. Because at our ICS AI Update on May 21, it became clear once again: Generative AI does not unfold its true benefit in a battle of tools, but rather when it is embedded into the real challenges of export companies.<\/p>\n

Together with ICS expert Iris Edler-Stiegler<\/strong>, Lackner took the more than 60 participants on a practical journey through the latest developments \u2013 from smart assistants and deep research to autonomous agents that automate research, market analysis, and follow-ups.<\/p>\n

Using the example of a potential organic cereal bar export to Japan, it was impressively demonstrated how AI can support market analysis at the touch of a button, intercultural training, and trade fair preparations. Concepts such as RAG (Retrieval Augmented Generation)<\/strong> were also used, allowing company data to be intelligently integrated into AI workflows.<\/p>\n

The use cases shown provided many aha moments \u2013 as confirmed by the feedback from participants: “I found the great practical examples particularly exciting. The fast-paced presentation style and the large amount of information also suited me perfectly!”, “Very good presentation with a lot of important practice-related content. Top!”, or “Very helpful for making everyday life more efficient!”.<\/p>\n<\/div>\n <\/div>\n\n <\/div>\n <\/div>\n <\/div>\n

\n
\n
\n
\n \n \"\" <\/a>\n <\/div>\n
\n \n \"\" <\/a>\n <\/div>\n
\n \n \"\" <\/a>\n <\/div>\n <\/div>\n <\/div>\n <\/div>\n
\n
\n
\n
\n \n

Another central topic was the Human in the Loop<\/strong> principle & the risks of so-called hallucinations<\/strong> \u2013 i.e., incorrect or misleading AI answers. After all, AI can automate many things, but especially in export, where cultural nuances and legal precision are crucial, humans remain indispensable as a verifying instance. All the more shocking are the results of a recent study by EY<\/a>, which states that only a quarter of respondents from Germany double-check AI-generated content, placing our neighbors even below the global average of 31%.<\/p>\n

Also particularly exciting: alternatives to cloud-based usage. Keyword On-Premise AI<\/strong> \u2013 systems that are operated locally and allow for maximum control over sensitive company data. While higher initial investments are often a barrier here, they bear no relation to data security and independence from providers and internet connections.<\/p>\n

Where will the journey lead? Lothar Lackner remains deliberately vague \u2013 but one thing is certain for him: “We are no longer looking for the one tool that can do everything \u2013 instead, we define business processes for which specialized tools can make our work more efficient.”<\/p>\n

Recommended for this are the latest releases from Microsoft and Google, practically summarized by CNET in 15-minute clips each:<\/p>\n