Most arguments about an estimate are really arguments about something else.
I spent years on the sales side, where a high number feels like an obstacle someone put in my way. Sitting closer to the people who do the building, I see it differently. When the person estimating doesn't know something, the safe answer is more hours. Nobody is being difficult. A vague request gets a padded reply, and that is a sensible reaction to being asked to promise something you can't see.
So when a number looks too big, I have stopped asking for it to be smaller. I ask what it assumes. How many things are we connecting? Who provides the documentation? What happens if access arrives two weeks late? Usually a few of those answers are "we don't know", and each one is quietly carrying a buffer.
What has worked best for me is a range with reasons attached. The low end is what the work costs if certain things are true: the customer sends the files early, the other system behaves as described, the right person is available to answer questions. The high end is what it costs if they are not. Written that way, the customer can see which parts of the price are in their own hands.
It also changes my job before the estimate is ever requested. If I bring back a loose description, I have ordered the padding myself. A handful of sharper questions in the first conversation does more for the price than any negotiation afterward.
I am still wary of a number that shows up before anyone has looked at the work. It is usually a budget, or a hope, or a figure that fit someone's approval form. It can be a fine starting point. I just don't want to treat it as a measurement.
I don't always get this right. I still catch myself pushing on the total when I should be asking what sits underneath it.