Technology Record - Issue 33: Summer 2024

103 Duke Energy has saved over $250 million with predictive models by AVEVA INDUSTRIALS & MANUFACTURING for optimal operational efficiency and sustainability gains. Later in 2024, AVEVA will introduce its industrial AI assistant to CONNECT. This patent-pending solution is based on generative AI and a large language model (LLM) that is specifically primed for industries to help customers intuitively locate and analyse data for top performance. Users can also ask complex questions, such as why their factory is operating at lower production levels today compared to last month. The industrial AI assistant breaks down that question into component parts and then queries various data sources to identify the root cause and provide the user with a consolidated answer. How will AVEVA’s industrial AI assistant support businesses and enable them to become more efficient and sustainable? Our industrial AI assistant combines an LLM, generative AI and AVEVA’s patent-pending technology to analyse customer data, find issues and help solve problems more quickly. Users can now ask natural language questions with minimal setup required. They can also ask objective-driven questions, which means more complex, multi-step requests can be completed in one single action. This can be automatically combined with other types of AI such as predictive analytics to detect asset and process issues early, allowing industries to quickly make their operation more efficient and thereby reduce fuel consumption and greenhouse gas emissions. With the support of an intuitive AI assistant interface, industrial teams can take advantage of LLM technology that leverages AVEVA’s more than 50 years of industry expertise to meet their desired goals. What does the future hold for industryprimed LLMs? Today, industrial LLMs are humanising AI so users can get more value from their data. This capability is expected to evolve and grow, which will incrementally and exponentially help our customers get value from everything we do and software functionality. Photo: Duke Energy