How Data Fabrics Contextualize AI Data – Podcast Transcript
Data is the core of a good AI solution and, while data is plentiful in enterprise and industrial settings, accessing and utilizing that data for AI training isn’t so easy. In industry, data sources are often fragmented and, even if they can be easily accessed, lack the critical context needed to support robust AI usage. This is where data fabrics shine, offering a single, unified way of accessing and contextualizing all types of enterprise data.
In this episode, host Spencer Acain is joined by Tobias Malbretcht, Head of Product Management, Data & AI at Siemens Digital Industries to examine the importance of data fabrics in the expanding role of AI in industry.
Check out the full episode here or keep reading for a transcript of that conversation.
Spencer Acain: Hello and welcome to the Future Ready Podcast’s series on artificial intelligence. I’m your host, Spencer Acain. In this series, we explore a wide range of AI topics from all across Siemens and how they’re apply to different technologies. Today I’m joined by Tobias Malbrecht, VP of product management data and AI for Altair, now a Siemens company. Welcome, Tobias.
Tobias Malbrecht: Hey Spencer, great to be here.
Spencer Acain: Great. Well, before we get into our main topic, can you give a brief introduction? Who you are and your background and your work with AI and Siemens.
Tobias Malbrecht: Yeah, absolutely. Yeah, hi, my name is Tobias. I’m very pleased to be here. So yeah, ever since I studied computer science and business administration at the technical university of Dortmund, I worked with data and AI. So shortly after my studies, I joined a startup company called Rapid Miner in 2008. And effectively, Rapid Miner was built on the promise to make data science and machine learning simple for people who are not trained data scientists. So obviously a lot of data science is done by experts but Rapid Miner Minor really aimed to put data science and machine learning into the tools of people who do not have a data background by training, so people like process engineers or technical engineers or mechanical engineers, and yeah, basically fast forward then a couple of years until 2022,there Rapid Miner has been acquired by a company called Altair, which just has been acquired by Siemens. And Altair obviously was founded way earlier in 1985, and it was a leading simulation company, but they also invested into high performance computing and into data science and data analytics. So and as Rapid Miner, we joined Altair and basically become part of a portfolio of data analytics products. And now with the acquisition of Alter by ZMES in 2025, so last year, yeah, we became part of Siemens, and now I’m really excited to lead the product management of this data analytics portfolio, which now takes the Rapid Miner name as the overall brand name. So basically, yeah, Rapid Miner platform data analytics portfolio that is now part of Siemens.And so yeah, that’s a very exciting combination.
Spencer Acain: Yeah, it is this exciting combination. So you’ve talked about you just mentioned Rapid Miner a few times there, and well, can you tell them tell us what that is? You know, what are what is Rapid Miner? What is a data fabric and why is that important for AI?
Tobias Malbrecht: Yeah, so basically Rapid Miner is a data and AI platform. So what it does, it helps you to structure data and bring data together, and it helps you to build data science, machine learning and AI from data that ultimately facilitate decision making on yeah, decision making from data. So a data fabric is in in that sense important because data is typically distributed all across an enterprise. And for data analytics, that is obviously a problem and for AI even more so if you have a question, you obviously, if you want to answer that question, you need to tap into all the data that well basically helps you answer that question. But as this data is distributed, that’s a problem because for answering questions, then you need to bring data together, basically, right? And the kind of legacy approach to that is whenever you have a question, you then sit and look at which data you need and bring this data together to answer that specific question. Now, Rapid Miner flips this around basically. So Rapid Miner allows to provide an overlay and over the enterprise data and the distributed data sources that an enterprise has and connect all these data within a knowledge graph. And what that does is it provides a kind of one place where you access data. So there’s only one place that you need to go to when you have a question for analysis, right? And data fabrics do that basically, and specifically Rapid Miner offers a knowledge graph based. data fabric that connects all those data sources across an enterprise than to do analytics
Spencer Acain: So it sounds like instead of you know basically connecting all of your data every time you have a question your you have all the data connected and then when you go in with a question your kind of your entry point automatically has all that data connected to it like you ask a question and then it already knows all the different places it needs to go to look instead of having to you know go through everything and maybe miss something is that kind of kind of where we’re at with this?
Tobias Malbrecht: Yeah absolutely and that has a bunch of advantages so obviously it speeds up analytics and so it accelerates the process of getting from the question to the answer so we talk a lot a lot about in in the in to our customers about decision latency right so ultimately you want to reduce the time from having the question to getting the answer for something well obviously that’s required to gain competitive edge and re be really competitive in in today’s fast paced market so to reduce that time that decision latency is very important and Rapid Miner helps to reduce that time by having this kind of knowledge graph overlay over the data always ready for your questions that that you answer want to answer so you always have basically your data stylus broken down and all the data connected and ready at hand and at your disposal to ask questions is it’s really important for analytics it’s also very important for AI because AI obviously also needs to tap into all the data right don’t want to wait basically days or months or even years until data is integrated AI just instantaneously needs that access to all this data hence it’s so important that all the data stays connected and data silos do not exist across the enterprise landscape and the data landscape of an of an organization hence it’s so important basically to have data connected and to be able then to provide all this data to AI as context so what we basically describe this as this layer is it’s a context layer that offers data toa decision making entity whether it’s an algorithm or whether it’s a machine learning model or whether it’s an AI agent that can provide this data as context and that encompasses all the proprietary data of an organization or the proprietary knowledge of an organization.
Spencer Acain: It sounds like that would also help with training AI wouldn’t it you have the context you have all the data right there so you’ve kind of talked about in the context of using it with AI and with machine learning to answer questions but can this then kind of be turned around and then could you use this to help develop the AI models to begin with as like a source of data for that?
Tobias Malbrecht: Yes absolutely so Rapid Miner serves both aspects it serves the context layer so basically structuring connecting and structuring all the data and making it available to an intelligence layer for decision making but it also serves the intelligence layer so the rapid map platform provides AI and automation toolkit and it allows users to build data science machine learning and AI models that actually facilitate that decision making from the data
Spencer Acain: And that’s something that’s you’re helping to support with Rapid Miner right that all of that kind of functionality like when you’re for an enterprise looking to design or to train their own AI models?
Tobias Malbrecht: Yes absolutely, correct so when we talk about like how to tune improve and reduce the decision latency of an organization that that I talked about there are basically three main enablers that do this there’s a the context layer which again represents all the data that that an organization has and Rapid Miller provides a way to connect all this data and then provide this in a knowledge graph for simple access and easy access by the intelligence layer secondly it also provides this intelligence layer so all the automation and AI toolkit for implementing models and then also for operationalizing those models and then there are obviously integration points where then the decisions that this intelligence layer makes, whether it’s a machine learning model or an AI agent, that can integrate those decisions into any business processes. So it can obviously expose results to data visualization tools like Tableau or Click or anything. It can expose decision making via APIs so that it can be integrated with other technical systems. But it also supports a combination with Mendix as an example, which also is provided by Siemens as a low-code application development platform. So that basically decisions can be exposed to consumers through specialized apps that you build in Mendex. And this layer, this layer that that Mendix basically is in we talk about this as the action layer. So ultimately, there are three layers and enablers to reduce decision layers in an organization. There’s context layer through a data fabric. There is the AI and automation toolkit as an intelligence layer to help organizations build an AI factory and kind of speed up decision making. But then there’s also an action layer that helps with integrating decisions into business processes, and that’s then facilitated through a local application development platform like Mendex that we also have in Siemens.
Spencer Acain: No that sounds like a like a great like one, two, three layer that you have going there. It makes a lot of sense to structure it that way. So you mentioned connectors a couple of times as well. And is that how you’re getting you’re getting data into Rapid Miner? Is that is that like a really easy process? Does it take some time, like what kind of data can you all put into Rapid Miner? You kind of mentioned earlier, of course, all the data across an enterprise, but surely that would be a very difficult task with all the different kinds of information and all the different systems.
Tobias Malbrecht: Yes. I mean, obviously, data can be stored in very diverse systems, right? So there are like databases, there are data storages, data lakes, there are certain enterprise and business applications. So it can come in all various forms. And we try to be very agnostic and very portable and composable in the sense that Rapid Miner can tap into all different kinds of data sources, and that goes often via like standard APIs. So there are often like standard interfaces to pull data from certain databases or data sources, and where there are no such data sources or data interfaces existent, well, then it’s quite easy actually to add a new connector to a new data source. and we can obviously always do that with our customers when they have a certain special data source. And yeah, data can be really diverse. So obviously we include a lot of like standard connectors to various like standard data platforms, whether it’s like a cloud data warehouse like Snowflake or data platform like Databricks or any other like database Oracle MS SQL, you name it, but we also provide special connectors to various tools. So obviously, if we speak within the Siemens ecosystems, there are various like tools that support building products and managing the product lifecycle, like a team center as a product lifecycle management tool, or like an op center as a manufacturing execution system, or like an insights hub as an IoT system. Um, and we obviously connect to all of such systems to make it very easy to tap into data that an organization has, and that is very important for decision making, especially automated decision making, as you would want to facilitate through AI and through A agents.
Spencer Acain: Thank you, Tobias, for those great insights on the applications of Rapid Miner and data fabrics within the AI space. But unfortunately, that is all the time we have for this episode. So tune in again next time as we continue this discussion on applications of data fabrics and AI and what that means for AI agents.
Siemens Digital Industries Software helps organizations of all sizes digitally transform using software, hardware and services from the Siemens Xcelerator business platform. Siemens’ software and the comprehensive digital twin enable companies to optimize their design, engineering and manufacturing processes to turn today’s ideas into the sustainable products of the future. From chips to entire systems, from product to process, across all industries. Siemens Digital Industries Software – Accelerating transformation.
