The "at least" part is what we need to work with at this point. We are working on learning how best to deflect such asteroids. We will probably learn some very useful things about Apophis when it goes by in 2029. And, it will go buy about every 17 years thereafter, giving us other opportunities to learn more and maybe do something to reduce risk of later passages. We may not even need to do something in 100 years, depending on what future observations allow us to predict.
It is hard to realistically predict technological progress, but it is at least safe to say that we are currently rapidly increasing our space capabilities. SO, within the next 100 years, it seems likely that we will have technological infrastructure available to change Apophis' orbit in a much more reliably safe manner than anything we could attempt in 2029.
What we probably are going to want to do is to slow its orbital speed around the Sun when it is at its closest approach to the Sun. That could get it's path completely inside Earth's orbit, so there would be no chance for a collision. But, how to do that is the question. Just hitting it with an impactor like DART would probably knock enough stuff loose that some of it could have a high probability of hitting Earth on the next pass. That is not what we want. Maybe we will want to put a solar sail on it that is angled 45 degrees from the Sun direction to bounce light towards its forward direction, very slowly reducing its orbital speed. But, to do that, we would need to deal with Apophis' rotation in some way. Or, maybe we need to capture it in what amounts to a large net, with a steerable rocket motor attached that could both stop its rotation and then slow its orbital speed. We can't do either right now, but, hopefully, we will learn how to do such things, or maybe better things, within 100 years.