Research programme
Modelling how entrepreneurs regulate cognition when interacting with AI
ICON builds on entrepreneurial cognition, entrepreneurship–AI research and social cognition theory to reconstruct metacognitive regulation as a process, and then measure its key dimensions with a purpose-built, cross-culturally validated instrument.
Context
Background
Entrepreneurial cognition research is vital to understanding how entrepreneurs mobilise cognitive resources and take action under uncertainty. Metacognition (the awareness and control of one’s own cognitive processing) has recently received growing interest in entrepreneurial scholarship.
At the same time, the “AI-fication” of the field is opening entirely new debates. AI is becoming one of the key actors shaping human agency in entrepreneurship, raising questions about the cognitive scaffolding of entrepreneurial thinking. AI is no longer viewed merely as a supportive tool, but as a cognitive copilot that augments entrepreneurs’ cognitive resources and influences how they regulate their thinking and make decisions.
The discussion also points to new cognitive challenges, especially the difficulty of evaluating AI output and deciding when and to what extent to rely on it. Such challenges make the calibration of trust in AI a central issue in entrepreneur–AI interaction.
Objectives
From process model to entrepreneur profiles
The main objective is to model how entrepreneurs’ interactions with generative AI shape their metacognitive regulation, and to identify and operationalise the key dimensions of that regulation and their patterns of variation across entrepreneurs. Four specific objectives follow.
-
RO 1
Identify and model the process
Identify how entrepreneurs’ interactions with generative AI shape their metacognitive regulation, and develop a process model of metacognitive regulation in digitally scaffolded entrepreneurial cognition. Addressed by the qualitative study, which models regulation as a broader process of monitoring and control within the entrepreneur–AI interaction.
-
RO 2
Define the key dimensions
Identify and conceptually define the key dimensions of metacognitive regulation in entrepreneurs’ interactions with generative AI, and distinguish them from related constructs. Cross-case analysis yields abductively derived patterns, leading to dimensions conceptualised as continua.
-
RO 3
Develop and validate the instrument
Develop and validate a measurement instrument for assessing key dimensions of metacognitive regulation. Item creation relies on the dimensions identified in RO2 together with existing validated scales.
-
RO 4
Identify profiles and test decision quality
Identify distinct profiles of entrepreneurs based on the dimensions of metacognitive regulation, and test for differences between profiles in the quality of entrepreneurial decisions.
Population
Knowledge-intensive SMEs
The research population consists of business owners and directors of SMEs operating in Croatia and Norway with primary activity in knowledge-intensive services.
SMEs are defined following the European Commission Recommendation 2003/361 of 6 May 2003, and knowledge-intensive services follow the non-public KIS economic activity sectors defined in the Eurostat glossary of knowledge-intensive services. Entrepreneurs are defined as individuals who own or manage a business and hold decision-making authority.
KIS contexts rely on intensive information processing and ambiguity, with a strong predisposition toward AI use.
Own calculations from the Eurostat database.
Design
Methodology
Two studies, one sequential design
A qualitative study reconstructs the process and develops the dimensional framework; a quantitative study validates its measurement and identifies empirically distinguishable entrepreneur profiles.
-
Study 1 · Qualitative
Task-based interviews with think-aloud protocol
-
Study 2 · Quantitative
Instrument development and cross-cultural survey
Foundations
Preliminary grounding
The planned objectives and design are novel and original, without prior empirical results, but they are built on established approaches that support their feasibility and relevance.
- Conceptually, the study draws on entrepreneurial cognition literature and its processual perspective, especially the situated metacognitive model of the entrepreneurial mindset.
- Methodologically, it relies on established approaches from cognitive psychology, including task-based and think-aloud protocols, and on prior work developing context-specific metacognitive measurement instruments.
- In entrepreneurship, task approaches have already been applied in the study of entrepreneurial cognition.
- In human–AI interaction, existing literature supports the feasibility of studying cognition and trust in AI.
ICON combines these trajectories in a novel way to examine entrepreneurial metacognition in AI-scaffolded decision-making, an area that remains underexplored.
Next
