![\"Image](\"https:\/\/blogs.sw.siemens.com\/wp-content\/uploads\/sites\/19\/2026\/04\/590-160775-keyvisual-overall-4zu3-rgb-original_original-1-1024x768.jpg\")
Is data the new oil? Only when it is contextualized<\/h2>\n\n\n\n
The metaphor to make data the new oil offers CEOs a great visual. A process engineer digs into a manufacturing facility\u2019s data lake. Using product lifecycle management (PLM) tools, they excavate the relevant information from the irrelevant and refine what is left into hidden opportunities and optimizations.<\/p>\n\n\n\n
Or perhaps a product manager extracts all the information from a sea of customer data. They use a consumer relationship management (CRM) tool to drill into their customer\u2019s preferences, use cases and feedback to pump out a new product better suited to the target audience.<\/p>\n\n\n\n
The key similarity between these two examples is that a human used a specialized data management tool. These tools contextualized the data and offered workflows to lead the user toward their data-driven insights. But the user\u2019s human insights were still needed as part of the equation. PLM and CRM are not the only examples of industry tools like this, enterprise resource management (ERP), enterprise asset management (EAM) and many others exist.<\/p>\n\n\n\n
Another lesson here is that the process of extracting meaningful, correct insights from vast swaths of uncontextualized data is not as easy as it sounds. Going back to the make data the new oil metaphor, humans know what to do with oil when we find it; we sell it, burn it or make it into products. We have the context clues available to make decisions on how the oil should be used.<\/p>\n\n\n\n
The same can not always be said about data. When we stumble upon a random vast database, it is often unclear what can, or should, be done with it. This leads to how data accumulates in forgotten records and data lakes. Data is a resource, like oil, but it is also a common afterthought: \u201cjust submit the report and we\u2019ll figure out how to use it later.\u201d The problem with this strategy is that without context, data is little more than a collection of numbers. And the longer data is shelved, the more context is lost to obscurity.<\/p>\n\n\n\n
This is another benefit of tools like PLM, CRM, ERP and EAM. Instead of data being based on reports hidden in physical or digital file folders, it is organized and contextualized with a product-, customer-, process- and\/or asset-centric frame of reference.<\/p>\n\n\n\n
How to make data valuable in an AI-based future<\/h2>\n\n\n\n
It might seem to CEOs that by plugging all the data from these above-mentioned tools into a machine learning algorithm would yield results. It is contextualized data, after all. That is what we need to make data the new oil. Additionally, many of these tools have similarities or even data overlap, so it should be simple for AI to understand how they interconnect, right?<\/p>\n\n\n\n
Again, it is not that easy. These are disparate systems. The relationships, links and structures between the data within them will not be clear to AI unless it is given the context it needs to understand.<\/p>\n\n\n\n
Theoretically, an AI models could learn from all the data between these tools and come up with connections and insights that would be hard for a human to discover using one or more of these tools in conjunction. But, without humans that understand the links, overlaps and connections between these tools, AI has no context.<\/p>\n\n\n\n
So, what would happen if AI were trained on these data siloes without the context it needs to understand their relationships? It would force the AI to make assumptions as to what those connections may be. And when bots make guesses, hallucinations are never far behind.<\/p>\n\n\n\n
The solution is that, like the data management tools, AI still needs human guidance. Humans must connect the dots for AI models using classification tools called ontologies: webs of data connected into a fabric that define assets, processes, products and customers holistically. And building these ontologies is not easy\u2014or instant. It is this realization that has led to so many AI projects to go over budget to only produce buggy, hallucinogenic models.<\/p>\n\n\n\n
Put simply, to make data the new oil in an AI setting, humans must tell the algorithms the relationships between disparate data silos\u2014like PLM, CRM, EAM and ERP systems. Only then can AI begin to make useful insights or properly answer questions like: \u201cWhat are the financial, sales, workforce and inventory implications of shutting down a mixer for a week, this July, for preventative maintenance?\u201d To make this possible, CEOs must understand the engineering work, time, efforts and costs associated with creating these ontologies\u2014those that don\u2019t are hallucinating.<\/p>\n\n\n\n