Artificial intelligence (AI) has increasingly been recognized for its ability to support financial professionals in providing their services. It can virtually aid in decision-making across various services, from asset management to investment recommendations. However, AI shows the most promise in detecting suspicious transactions and client activities within the context of anti-money laundering (AML). In this capacity, AI can quickly analyse large volumes of data to assess whether certain behaviours or transactions pose a risk of money laundering. Despite these technological advancements, the use of AI systems must be appropriately regulated. When drafting suspicious transaction reports (STR) and suspicious activity reports (SAR), AI can assist compliance officers by enhancing their investigations and improving overall detection for financial professionals. However, to what extent can compliance officers rely on AI to base these reports? And how can AI outputs be integrated into such reports without exceeding their prudential obligations under AML? These questions are crucial for establishing a compliant framework for AI in detecting suspicious transactions and activities.