AI tools can generate data analysis that looks polished, comprehensive, and credible. What they cannot do is own the decision that follows from it. That responsibility stays with the leader presenting the analysis, defending the assumptions behind it, and making the case for a specific course of action. When leaders treat AI-generated analysis as a finished product rather than a starting point, they are not making faster decisions. They are making less accountable ones, and creating real professional risk in the process. Data reasoning is a leadership competency, and it has layers that most organizations are not yet developing intentionally: interrogating assumptions, understanding data quality, reading the audience, and building the narrative that connects analysis to action. AI changes the tools available for that work. It does not change who owns it.

There is a version of the AI productivity conversation that goes roughly like this: AI can now handle the analytical work that used to take hours, so leaders can spend less time building models and more time making decisions. The implication is that the hard part, the data work, has been automated away, and what remains is judgment.

That framing is partially right and significantly dangerous. AI tools have genuinely changed what is possible in data analysis. A leader who once needed a financial analyst and several days can now get a formatted projection in minutes. The speed is real. The problem is that speed without interrogation is not productivity. It is risk transfer, from the tool to the person presenting the output.

When the analysis is wrong, or built on assumptions that do not reflect reality, the AI does not stand in front of the room and absorb the consequences. The leader does.

The Spreadsheet That Looked Beautiful

Consider what this looks like in practice. A leader uses an AI tool to build a long-term financial projection. The output is well-formatted, comprehensive, and visually credible. It looks like something a skilled analyst spent time on. But when the leader goes through it cell by cell, which takes time and requires genuine financial literacy, they find that the tool has made a series of assumptions that are incorrect for their specific context. Some of the calculations are wrong. The projection, presented as-is, would not survive the first round of questions from anyone who knew the business.

This is not an edge case. It is a predictable feature of how AI tools work. They generate outputs based on patterns and general assumptions. They do not know your organization’s specific history, your data’s quirks, or the context that would cause a knowledgeable person to flag a number as implausible. That knowledge lives with you. The tool does not have it, and it cannot ask for it.

The leader who catches those errors before presenting has done the real work of data reasoning. The leader who does not catch them will find out in the room, in front of the people they were trying to persuade, at exactly the wrong moment.

Data Quality Is a Precondition, Not a Given

There is a layer underneath the assumption problem that is worth naming explicitly: the quality of the data going into the analysis. AI tools produce outputs that are only as reliable as the inputs they work with, and in most organizations, the underlying data is messier than anyone wants to admit.

Data entry conventions change over time. Systems get migrated and fields get mapped inconsistently. Definitions shift without documentation. A column that means one thing in year one means something slightly different in year three, and nothing in the data itself flags that change. When an AI tool runs analysis on that data, it treats the inconsistency as information rather than noise. The output looks clean. The underlying problem is invisible unless someone who knows the data is paying attention.

Understanding your data well enough to know where it is reliable and where it is not is not a technical skill reserved for analysts. It is a leadership responsibility for anyone who uses data to make or support decisions. AI tools do not change that responsibility – they actually make it more consequential, because the outputs they generate look authoritative even when the inputs are not.

The Narrative Is Still Yours

Even when the analysis is sound and the data is clean, there is a second layer of work that AI cannot do: building the narrative that connects the analysis to a decision and persuades a specific audience to act on it.

Presenting data is not the same as making a case. A case requires knowing who is in the room, what their concerns are, what objections they are likely to raise, and how to answer those objections in a way that moves them toward a decision rather than deeper into doubt. It requires understanding what action you are trying to achieve and structuring everything, the data, the framing, the sequencing of information, around that outcome.

An AI tool can help you generate analysis. It can even help you anticipate questions if you prompt it well. But it does not know your CFO’s specific concerns about this quarter’s numbers; it does not know that one board member will push back on any projection that assumes flat headcount; it does not know the organizational history that makes a particular framing land as credible or as tone-deaf. That knowledge is yours. The persuasion is yours. The outcome, whatever it is, is yours.

Data Reasoning as a Leadership Competency

The OGC Leadership Competency Wheel includes Data Reasoning and Decision Making as a core competency domain, and the AI conversation is one of the clearest illustrations of why that framing matters. Treating data reasoning as a competency rather than a technical skill changes what organizations develop for and how they develop it.

A competency lens sees the layers: the ability to interrogate assumptions, to understand data provenance and quality, to construct a coherent narrative from analysis, to read an audience and anticipate their objections, and to connect all of it to a clear decision and a defensible recommendation. These are not skills that emerge naturally from access to better tools. They are skills that have to be developed intentionally, in the same way that any other leadership competency is developed: through structured practice, feedback, and progressively higher-stakes application.

The organizations that will use AI well are not the ones that give their leaders the best tools. They are the ones that develop leaders who know how to own the work those tools produce. That is a development investment, not a technology one. And it is one that most organizations have not yet made deliberately.

AI is changing the analytical landscape faster than most leadership development programs are keeping up with. The leaders who will navigate that gap well are the ones who understand that the tool is not a substitute for judgment. It is an input to it.

If your organization is thinking about how to develop data reasoning as a leadership competency rather than a technical skill, we are happy to talk through what that looks like in practice. Learn more about the OGC Leadership Competency Wheel or reach out to start that conversation.  

Have a specific question about how this applies to your organization? Our FAQ page covers common questions on talent systems, leadership development, and engagement data.