INTELLIGENCE IN REAL LIFE
Work — Matching people, cognitive profiles, tasks and conditions
Work performance depends not only on how much cognitive ability a person has, but on whether their particular strengths and limitations fit the tasks they are asked to perform and the conditions in which they must perform them.
The central problem is cognitive fit
General cognitive ability is a replicable predictor of job performance, but contemporary meta-analytic estimates are more modest than some influential older estimates and vary with criterion and job characteristics. Specific abilities can also matter for particular forms of work. Organisations therefore should not treat a single general score as a complete description of occupational capability: they employ someone to perform a bundle of tasks under particular conditions.
A person may be exceptional at analytical reasoning and poor at organisation. Another may be highly organised and reliable but struggle with rapid novel problem solving. A third may perform brilliantly in sustained individual work and poorly in a role built around constant interruption. Treating all of these as a single 'ability' problem loses the information needed to design good work.
Jobs should be decomposed into cognitive demands
Job titles are too coarse. Work can be mapped into tasks requiring different combinations of reasoning, memory, planning, processing speed, attention, verbal communication, social cognition, visuospatial skill and acquired expertise. Tasks also differ in novelty, ambiguity, interruption, time pressure and consequence of error.
Once those demands are visible, organisations can ask whether they are essential. Does this analyst genuinely need to coordinate ten calendars? Does this manager need rapid mental arithmetic? Does a highly repetitive administrative requirement belong with the person doing the most complex conceptual work? Some demands are central; others are historical accidents of workflow.
When fit is poor, there is more than one intervention
If someone struggles with an essential capability, training may be appropriate. If the difficulty is stable but peripheral to the person's core contribution, support or task redistribution may be better. If everyone struggles, the workflow may be the problem.
Tools can externalise memory and organisation. Teams can distribute complementary strengths. Interfaces can reduce interruption. Roles can be redesigned. Accommodations can change timing or modality. None of these automatically lowers standards; they can move cognitive effort toward the work that actually creates value.
State can turn a good fit into a bad one
Workloads fluctuate. Sleep loss, illness, anxiety, burnout, caregiving pressure and acute stress can reduce usable cognitive resources. Time pressure can make a normally manageable task dependent on speed. Constant interruption can make a role depend more heavily on working memory and task switching than its substantive purpose requires.
This means occupational performance should not always be interpreted as a stable signal about the worker. Sometimes it is a signal about current load or work design.
Selection and inclusion need the same cognitive analysis
Assessment can help organisations identify people likely to succeed, but selection methods should correspond to real task requirements. If an assessment adds cognitive demands that the job does not require, it can screen out capable people for the wrong reason.
The same analysis supports inclusion. Disability accommodation becomes more precise when the organisation understands which function is impaired, which task is affected, whether that demand is essential, and what alternative route preserves the outcome.
AI can rewrite person-task fit
Generative AI can supply structure, drafting, retrieval, translation, coding support and planning. A cognitive weakness that once constrained a role may become supportable. At the same time, tasks that previously differentiated workers may disappear or change.
This creates new questions about assisted performance. If AI allows a worker to perform a task well, that can be a legitimate expansion of effective capability. But organisations also need to know what happens when the system is wrong or unavailable and whether the worker has enough domain knowledge to supervise it.
What HIW should investigate
Which cognitive profiles fit which tasks, not merely which occupations? Where do organisations accidentally create demands unrelated to the work's purpose? When should a mismatch be addressed through training, accommodation, task redesign, team composition or technology? How do stress and overload change fit over time? Can AI widen access to work without creating fragile dependence?
A mature science of intelligence at work should help people contribute where their abilities are most valuable and receive support where a nonessential weakness would otherwise obscure those abilities.
Selected evidence and further reading
- Hunter — Cognitive ability, job knowledge and job performance
- Sackett et al. (2024) — A contemporary look at general cognitive ability and job performance
- Salgado & Moscoso (2019) — Meta-analysis of general mental ability validity for five performance criteria
- Hambrick et al. (2024) — Cognitive ability and job-specific performance across experience levels
- Lang & Kell (2019) — Specific abilities in the workplace: more important than g?
OPEN TO REVIEW AND CONTRIBUTION
Help improve the work.
Questions, expertise, references, corrections, lived experience and research collaboration can help make this work clearer, stronger and more useful.