Strategy Is What Your Organization Actually Pays Attention To

There is a large organization somewhere right now with a detailed AI governance policy. It is probably twenty or thirty pages long. It covers data privacy, model transparency, bias assessment, human oversight, and escalation procedures. Nobody with actual authority to act on it has read it this week. Nobody will read it next week either. Not because they are negligent. Because they have forty other things competing for the same two hours.

Ocasio (1997) made a specific argument about this situation, though he was not writing about AI governance. His attention-based view of the firm starts from the observation that firm behavior depends on what executives pay attention to, not on what the firm intends, plans, or formally decides. The argument has three premises. First, the focus of attention: executives direct effort, time, and cognitive resources to a subset of available cues and problems. Second, situated attention: what executives notice is structured by the situations they are in, the communications they receive, and the organizational contexts they inhabit. Third, the structural distribution of attention: organizational structures, communication channels, and decision-making procedures determine whose concerns get raised, whose get filtered out, and which problems ever reach people who can act on them.

This is uncomfortable because it implies that good ideas, good analyses, and good reports are not sufficient. A vulnerability report sitting in a security team's ticket queue has not reached the executives who could authorize remediation. A digital transformation strategy document that senior leadership signed off on but has not revisited in fourteen months is not driving any behavior. The signed document is not the strategy. The attention is the strategy.

Simon (1947) laid the groundwork for this. His bounded rationality concept established that decision makers are not omniscient rational agents but human beings with limited information, limited cognitive capacity, and limited time. Organizations do not optimize. They satisfice: they find solutions that are good enough given the constraints. The attention-based view takes bounded rationality a step further and asks: given that attention is genuinely scarce, what determines where it goes? The answer is that organizational structures and communication channels determine that. What gets on the agenda gets decided. What never gets on the agenda never gets decided, regardless of its objective importance.

I find the IS implications of this view more interesting than they usually get credit for. Take IT security incidents. The common post-mortem finding is something like: the vulnerability was known for months before it was exploited. Patches were available. The report had been filed. Nothing happened. The attention-based view would say the report was in the queue, but getting into the queue is not the same as getting on the agenda. The organizational structures and communication channels that exist in most IT organizations create a separation between the people who identify problems and the people who can authorize solutions. A report filed does not become attention received, and attention received does not become action taken unless it reaches someone with both attention and authority at the same moment.

The same pattern shows up in digital transformation. Organizations that succeed at transformation show a specific pattern, in my reading: sustained executive attention, not just a launch event. The initiative gets on the agenda at the leadership level and stays there through the hard middle period when the novelty is gone and the costs are becoming visible. Organizations that fail at transformation often have perfectly good strategies, adequate budgets, and committed teams. What they lack is sustained attention from people with the organizational authority to make the structural changes the transformation requires. The strategy document exists. The attention does not sustain long enough to change the structures.

AI governance is the clearest current case. Most large organizations now have AI governance frameworks. They have committees and checklists and review processes. What they often do not have is anyone in that governance structure whose primary job is AI governance and who has the organizational access to make those governance processes actually shape what product teams build and deploy. The governance framework exists at the layer of documents and procedures. The attention and authority needed to make it real are distributed to people whose primary attention is somewhere else.

Ocasio's structural distribution of attention is the piece I find most useful for IS research. He is not just saying that attention is scarce, which everyone knows. He is saying that the scarcity is structured by the organization. Different organizational designs distribute attention differently. A firm that creates a dedicated chief AI officer role with reporting access to the board is making a structural choice about where AI-related attention lives. A firm that assigns AI governance as an additional responsibility for the CISO or the legal team is making a different structural choice, one that puts AI governance in competition with everything else on those people's agendas. The outcomes of those two structural choices will be systematically different, not because the people are different, but because the structural distribution of attention is different.

The platform governance version of this is interesting too. Regulators who try to govern platform companies face a fundamental attention asymmetry. The platform company has hundreds of people whose full-time job is understanding the regulatory environment and optimizing for what regulators are watching. The regulatory agency has a smaller team watching a much larger information space. Platform companies are not just responding to regulatory attention. They are, in some cases, actively shaping what regulators pay attention to. That is organizational strategy operating at the attention level.

My read is that the attention-based view is one of the more honest theories about how organizations actually work. It does not pretend that good intentions translate into good outcomes through some rational mechanism. It asks the harder question: who actually pays attention to this, and what are the structures that determine whether their attention reaches anyone who can act? For IS researchers, that question is uncomfortable. It implies that our findings, our reports, and our recommendations do not matter much unless someone with attention and authority reads them at the right moment and has a structural channel to act. The quality of the research is necessary but not sufficient. The structural distribution of attention is the other half of the equation.