The "to be sure" tactic is usually a three part structure, consisting of an Evidence part, a Deflection Part, and a Conclusion part. The point of the deflection part is to defuse relevant context by making the reader feel "I've been heard."
The first part of the argument displays evidence (often presented as a Gish Gallop). For example, an argument against affirmative action in college admissions will point out its effects on Asian students. It will point to low graduation rates for Black students. It will say that Black students are ill-served after enrollment. And so on.
In readers sympathetic to, or neutral on, affirmative action, this will likely build a feeling the author is being unfair. What about X? What about Y? This builds up a certain tension that the author would be wise to resolve.
To do that, the author rolls out a section that acknowledges these likely concerns. "To be sure, Black students have likely received a poorer secondary education than the average student, and are considerably more likely to the first in their family to go to college."
Now that the tension has been resolved, the author is free to add more evidence and suggest a conclusion. The Tell that the "To be sure" section is in bad faith is that **the acknowledged content plays no part in the solution** and is probably not mentioned again.
In the case of affirmative action, the proposed solution will *not* include spending money improving public schools, or funding better mentoring for first-in-their-family college-goers.
See also Assume a Can Opener.