Not long ago (remember?), digitalization was something for nerds and unpaid interns. These days, even facility managers are talking over lunch about using AI to analyze floor plans. Progress, right?
And yet, when it comes to the data companies juggle every day, the pain points sound suspiciously old-fashioned: too much data, too many formats, too many systems—and no time. ERP here, Excel there, a DMS from a decade ago, a bit of technical building equipment, maybe some IoT, and somewhere in the mix a CDE platform no one knows how to use. But hey, at least it’s “digital.”
Same dashboard, different day
To stay on top of the data circus, companies are pouring millions into business intelligence systems—tools that promise to turn chaos into insights. But just try making a quick connection between data points. Say: “What were the utility costs last year for all sites needing renovation?” Suddenly, you’re waiting for the next BI training or hoping your colleague Tom isn’t too busy to help.
The promise was supposed to be different: less Excel hell, more meaningful insights. Instead? License fees, integration headaches, outdated data. And worst of all? Many systems are so rigid they react to change like a picky cat confronted with new organic litter.
The great cloud temptation
No wonder eyes are turning skyward—to the cloud, now home to the buzzword of the decade: Artificial Intelligence. “AI as a Service,” “Prompt Intelligence,” “Enterprise Copilots.” It all sounds fancy—until you start wondering: where does our data actually go when we talk to AI?
Many AI systems run on servers in the US, China, or other places where data protection is more of a nuisance than a norm. Feeding sensitive company data into those systems often means losing control—without even realizing it. And this isn’t just about planning interns’ vacation schedules—we’re talking trade secrets, contract terms, and construction budgets.
Charly, the data gladiator
AI can be incredibly useful—but only if it understands what it’s talking about and knows the context of your data. And we, the users, need to understand where that data is processed. Ideally, it stays with us. In-house. So what’s the solution?
You build an AI that doesn’t dump everything into the cloud. One that runs locally, respects data privacy, and still gets the job done. One that doesn’t pose as a digital know-it-all but acts like a smart colleague: quick, discreet, and surprisingly helpful.
That’s Charly, the chatbot. He’s the result of the Chat4BI research project, funded by the German state of Hesse and developed by speedikon FM AG and Innomatik AG. This AI runs completely locally. No data is transferred to third parties—and definitely not to any cloud. You get access to exactly the data you need right now—from blueprints and invoices to energy and IoT sensor data. And yes, Charly can also narrow it down to only the most up-to-date information.
Context is king
Unlike typical language models that sometimes make up facts (a politer way of saying that is that they “hallucinate”), Charly provides reliable answers—with source references. Want to know why Charly says heating costs have risen over the past three months? You’ll get the relevant reports delivered right alongside the answer.
Charly also understands graphical data. Reading blueprints? No problem. Interpreting floor plans? Easy. Deciphering handwritten notes? Sure thing. Charly’s like a chameleon that can spot the tiniest scribble from a kilometer away—nothing gets past him.
A new way to work with data
Instead of forcing users to click through five different systems, information goes straight to where it’s needed—to the user. All it takes is a simple chat: “Show me all heating costs from the past three months.” Just like that, data becomes a resource, not a burden. And decisions? Finally based on solid information—not “Patrick from Tech’s gut feeling” (sorry Patricks, you’re doing a great job!).
Why this matters
Because as a society, we need to learn how to use AI not naively, but responsibly. And we need solutions that reflect our European values: security, sovereignty, and individual rights.
So, go ahead and ask yourself:
Have you chatted with your data today? …and was anyone else reading too?
Picture: speedikon FM AG
