We propose just-in-time memoryless trust for crowdsourced IoT services. We leverage the characteristics of the IoT service environment to evaluate their trustworthiness. A novel framework is devised to assess a service's trust without relying on previous knowledge, i.e., memoryless trust. The framework exploits service-session-related data to offer a trust value valid only during the current session, i.e., just-in-time trust. Several experiments are conducted to assess the efficiency of the proposed framework.