Even Within an Organization There Is No Single AI Strategy

AI excels in tasks with clear rules and verifiable outcomes, such as coding or engineering, where feedback loops improve reliability. However, its performance drops sharply when tasks lack objective standards or relevant context, leading to unreliable or even dangerous outputs. Companies must therefore assess each task’s suitability for AI rather than assuming a one-size-fits-all approach will work. The key is to match AI’s strengths with the right kind of work to avoid costly mistakes.

When AI lacks training or feedback for a task, human expertise becomes critical to provide missing context, such as supplier reliability or customer exceptions. Organizations that capture tacit knowledge—through tools like head-mounted cameras or expert interviews—can give their AI a competitive edge. Preserving proprietary data and decision patterns is essential to maintain this advantage and prevent external models from replicating it.

In marketing and strategy, AI often produces competent but unoriginal output, risking convergence with competitors’ ideas. Humans must curate and refine AI-generated content to ensure it reflects the company’s unique identity and values. The role shifts from execution to oversight, focusing on what feels distinctive rather than merely competent, while using eventual outcomes to refine future AI inputs.

For genuinely novel decisions, such as acquisitions in unfamiliar markets, AI’s lack of relevant precedent makes it unreliable for predicting long-term outcomes. Human judgment remains indispensable in these cases, as AI cannot generalize effectively from sparse data. The challenge is to identify where AI can augment decision-making without overestimating its ability to handle ambiguity.