Guys Can the complexity of a flow affect the precision of the Query KB card ? As you can see my workflow has 2 flows here, the first one need to go through a bunch of nodes and AI tasks cards, the other just go straight to the final node (as shown). And when I tried to test with the same questions but through different flows, I notice that the 1st flow (heavy & complex one) has some problems with answer's accuracy as it give the wrong answer to my question unless I specifically tell it to answer the question precisely while the 2nd flow (lighter & shorter one) always give the right answer.