Background: AI hiring tools promise efficiency but often replicate historical bias embedded in training data. SMEs face unique hurdles as Swedish organisations rapidly adopt these systems, with limited resources, informal governance structures, vendor opacity, and mounting regulatory pressure under the EU AI Act. Organisational practice is understudied since current research concentrates on technical bias. This study examines how Swedish SME recruitment practitioners use AI in hiring and how they perceive, navigate, and implement fairness.

Purpose: This study examines how recruitment practitioners understand and implement fairness in practice, as well as how Swedish SMEs employ AI-based hiring tools. It addresses the discrepancy between organisational realities and fairness governance amid resource limitations, new regulatory demands, and technical fairness guidelines.

Method: This study employs a qualitative multiple-case study design with an abductive research approach. In total, 12 semi-structured interviews were conducted with Swedish recruitment practitioners across different industries.

Conclusion: The study shows that fairness in AI‑based hiring within Swedish SMEs is shaped primarily by practitioners' interpretations rather than by formal governance or regulatory engagement. Most participants view fairness as retaining human judgment and integrating multiple sources of evidence, which leads them to override or contextualise algorithmic outputs. Limited awareness of EU AI Act obligations and reliance on vendors create accountability gaps, highlighting the organisational challenges SMEs face as regulation approaches.