In this article, we delve into the world behind the scenes of digital transformation and AI-driven operations. Why is the pressure on IT managers so intense? How can business managers help alleviate these challenges? What does it take to truly strengthen your organization? Understanding and collaboration are more important than ever. Read on to discover what your IT department really needs to keep performing.
Holy grail
Let’s start with the drivers from the business side. These can be summarized briefly: AI and Generative AI (GenAI) represent an unprecedented leap forward. These technologies enhance efficiency by enabling automated processes to seamlessly anticipate rapidly changing circumstances. Business insights are supported in real time from an almost “infinitely broad” perspective. Additionally, customer experiences are improved through hyper-personalized interactions. GenAI takes this even further, offering creative solutions comparable to what an employee might “imagine.”
Tension field
Society expects organizations to respond instantly and “spot-on.” Organizations that embrace AI remain competitive. Those that do not risk falling behind in a digital economy where speed and intelligence are the winning factors.
We often hear how IT can be a bottleneck, for example, when implementing new tax legislation. With the modern business manager’s expectations, the pressure only grows, despite all the “IT-for-IT” efforts. And that’s not all.
CIOs and IT managers are already busy. Cybersecurity demands their full attention, alongside pressures like GDPR compliance and the EU AI Act, which will be enforced starting in 2026. The mandate for IT is simple: deliver an IT environment that allows for easy, controlled, and, above all, efficient deployment of AI.
IT management shifts the boundaries
Paradigms will shift to make the playing field “AI-ready.” This, of course, involves data for business operations—internally, with customers, and with suppliers. Cloud providers increasingly offer plug-and-play scenarios to deploy AI technology and integrate it with other IT environments. But the data? That’s up to IT managers to handle. And this is exactly where the boundaries need to shift fundamentally.
Earlier, the “infinitely broad” perspective was mentioned. While BI functions well with structured data, often SQL databases, AI requires well-prepared and often unstructured data, such as documents for training large language models (LLMs). The lack of expertise, tools, and skills to process and transition various forms of unstructured data from development to production smoothly is one of these shifting data paradigms. If IT cannot master this data processing and preparation, the potential of GenAI is lost. Business teams cannot enhance model performance because development cycles yield little innovation due to a lack of additional and/or better data sources. Cost-effective implementation then becomes more fiction than reality.
Even more intriguing is that AI, generative or otherwise, needs to consume a wide variety of data structures simultaneously. Streaming data and batch processing converge, each with its own demands on IT infrastructure, tied to a broad spectrum of data sources. Then there’s RAG (Retrieval Augmented Generation), which adds additional facts in the form of both structured and unstructured data to an LLM for answering a specific user query.
Beyond the IT complexity of managing these diverse data streams and structures, the suitability of these data sources is an increasing concern. Completeness, accuracy, and confidentiality of data form the foundation of every AI-generated response. This very puzzle often holds organizations back from broader adoption of GenAI solutions. And we haven’t even touched on potential biases in datasets. The primary concern typically revolves around the confidentiality of certain data elements and where this data ultimately resides. This, in turn, significantly undermines the business case for profitable AI applications.
Business management holds the key
Another paradigm shift in many organizations is that it’s not the IT department or database administrators (DBAs) but the business itself that holds the key to ensuring completeness and a fine-grained classification of data confidentiality. Without clear direction from the business, the broad automated use of data becomes impossible. Where data can be used, implementation proves extremely costly. Labor-intensive manual checks by DBAs and data engineers remain necessary before data can be utilized for AI purposes.
In the tension field described earlier, the solution, as always, lies in collaboration between business and IT. Here are a few keys that business management holds to ensure AI success. The business significantly influences the reliability, completeness, and timeliness of data from customers and suppliers. When making agreements, it’s the business that can demand transparency about the origin of data and whether it was collected with the right intent. And last but not least, it’s the business that, together with its stakeholders, ensures the provision of necessary metadata so that automated data processing can occur responsibly. For many organizations, this too will represent a shift in responsibility from IT to the business.
End well, all well
IT departments focus on building flexible, future-proof data and AI platforms. Business managers have to take responsibility for streamlining data management and creating clear agreements about data usage and quality guarantees. By jointly investing in knowledge and practices, organizations can unlock the full potential of AI.
So, want an AI future with fewer worries? Ensure your data is as reliable as your best friend, and treat your IT department as the star of your company (with coffee, compliments, and the occasional day off) rather than a burden. AI might be smart, but without good collaboration, it’s just an expensive calculator.
