Why AI Should Help Managers Focus — Not Simply Know More
Small and medium-sized enterprises are entering a new managerial reality: information is no longer scarce. Sales figures, customer feedback, competitor movements, economic indicators, supplier conditions, exchange rates, market trends, and AI-generated analyses can all reach a manager within minutes.
Yet having more information does not necessarily lead to better decisions.
In many cases, the real problem is knowing which information actually matters.
For an SME manager operating in a volatile market, attention is a limited resource. A broad economic report may be interesting, but it may have little relevance to a decision about next month’s inventory. By contrast, a sudden increase in customer enquiries, a change in supplier delivery times or a significant movement in replacement costs may directly affect that decision.
This distinction between information availability and decision relevance is becoming increasingly important in the age of artificial intelligence.
AI has dramatically reduced the cost and time required to generate analysis. Managers can now request forecasts, summaries, comparisons, and scenarios almost instantly. But this creates a paradox: when analysis becomes easier to produce, managers can end up receiving more outputs than they can meaningfully evaluate. The problem shifts from information scarcity to informational overload.
The solution may therefore not be another system that produces more information. It may be a system that helps determine what deserves attention before analysis begins.
A proposed approach known as the Relevance-Based AI-Assisted Decision Support Framework (RADSF) suggests reversing the traditional data-first model. Instead of collecting and analysing everything available and then asking what matters, the manager first defines the decision: What needs to be decided? Within what time horizon? Under which constraints? What outcome is most important?
Only then should AI help identify and prioritise the information connected to that decision.
This is a fundamental shift in the sequence of managerial reasoning: decision clarification comes first, information selection second, interpretation third, and action last. Prediction may be used within this process, but prediction is not treated as its central purpose.
The distinction matters because a sophisticated forecast is not automatically more valuable than a simple operational signal. For example, a recent change in supplier lead time may be more relevant to a purchasing decision than a complex macroeconomic forecast. Relevance is therefore contextual: information becomes important because of its relationship to a particular decision, time horizon and business constraint.
Consider an SME deciding whether to increase inventory for the coming month. Instead of presenting the manager with hundreds of pages of market analysis, an AI-supported system could prioritise recent sales velocity, confirmed customer enquiries, current inventory, supplier lead times and relevant currency movements. Broader economic information would not necessarily disappear; it would simply receive lower priority unless it becomes directly relevant.
The manager would then evaluate the prioritised signals, consider trade-offs and choose among feasible alternatives. If demand is increasing while supplier delivery times are lengthening, increasing inventory may become more reasonable. If demand remains uncertain and replacement costs are stable, a smaller order may be preferable. RADSF does not determine the answer; it improves the quality and focus of the information entering the decision process.
The framework also recognises that relevance is dynamic. A signal that appears unimportant today may become critical tomorrow when market conditions, the decision horizon or business constraints change. Relevance should therefore be reassessed rather than treated as a permanent characteristic of information.
This human-centred boundary is essential. AI can organise evidence, identify relationships, compare changes and rank signals, but it cannot fully capture tacit knowledge, unusual local circumstances, ethical considerations or the consequences of a decision that are not represented in available data. The manager must retain final interpretation and accountability.
For SMEs, this could become strategically significant. Smaller firms often operate with fewer analytical resources while remaining highly exposed to changes in demand, prices, supply chains and cash flow. Their competitive advantage may therefore depend not on analysing everything, but on identifying the few signals that can change a decision.
Responsible implementation also requires reliable sources, visible uncertainty, appropriate data governance and the possibility of human override. A relevance score should guide attention, not create an illusion of mathematical certainty. Poor data or a poorly defined managerial question cannot be corrected simply by adding AI.
The broader implication is important: the value of AI in business should not be measured only by how much information it can process or how many forecasts it can produce. In an environment of information abundance, value may increasingly come from knowing what not to place at the centre of attention.
AI should therefore not simply help businesses know more. It should help them focus better — while keeping judgment, accountability and responsibility firmly with the human decision-maker.
Dr. Sareh Goudarzi
Business Advisor, Business Development Specialist