Building a Document Chatbot Using Cortex AI at Siemens Energy

This transcript was automatically generated

Siemens Energy is a worldwide global player in the energy sector. We are an integrated company from energy production generation, bio transmission, to us, the households and the companies. We, as business, have a problem that has to be solved now. Sometimes our data is still literally on paper with valuable data in it.

So there was technology information which is necessary to to move a certain technology into the next level. And the only choice you had was, like, reading all the pages or create something that helps you to read the pages. Gen AI helps people to explore data which is literally sitting on paper. We have to digitalize it, and then we apply AI chatbots, especially with Cortex, on those datasets.

With the Snowflake, AI data cloud, we were really able to democratize data and to build, apps easily. The first use case is about those roughly eight hundred thousand pages that we now read, automatically or make them chatable. The purpose of this one was really to help those PhDs to find relevant information for design changes, cost changes. So this chatbot is now instead of providing very detailed knowledge about something, now this chatbot is used for a global thing, empowering people, training people.

And the benefit, of course, using Snowflake is that we manage all our data and apply AI with Cortex.

And then last not least, there is the user interface powered by Streamlit. This is, of course, from a security point of view, quite a benefit for us. We don't have any data floating around outside.

That is good. And what I'm really excited about is, the speed and the speed of innovation. The advantages using Cortex was, I mean, I literally was surprised how easy it is. Our first chatbot took us just a couple of weeks to make that happen.

And this was amazing, which is motivating me that we can, get more and more use cases tackled with this kind of Cortex approach. Knowledge in our company is literally sitting in the heads of people. So that means it's for us key, to push data democratization. We already have use cases in mind.

Next time we chat on sales documents, then we have ideas to chat on factory procedures to train people.

There there is a huge variety of use cases that we wanna tackle step by step. This is, at the moment, our strategy to make use of Snowflake and AI, generative AI, to explore those types of data.