As soon as a critic produces a criticism and indicates that there is a potential conflict, it has to decide if the criticism actually signals a conflict or not.
This is usually an easy task for the agent since it knows the reason why it produced that particular criticism. If the criticism was due to the violation of a hard constraint that tests some physical properties or user requirements, then there is probably a conflict. On the other hand, if the violated constraint that produced the criticism was just a preference of that agent, then there is no conflict and the design can go on. The criticism stays on the blackboard and the agents whose targets were criticized may take it into consideration if they wish.