Charge coolers can be done right, but it is complex... then again, so is adding a ton of volume and high-flow, lag-inducing charge-air plumbing for A2A intercooling.
Liquid being higher density than air has greater surface area contact, and is generally a bit more forgiving to plumb.
But the thermal exchange has to be right in two places, both for charge-air to liquid, and liquid to ambient air, as well as additional pumping for circulation.
I have heard that some charge coolers use the cold-side of the engine coolant circuit (after the radiator has cooled it, before the engine heats the coolant again) for charge cooling... but then if the engine runs hot, the radiator temperature drop may not be enough of to also be far enough below the intake charge temperature to pull that heat away, too.
Plus, if the heat exchangers in the intake tract are near the cylinder heads, not heat shielded and insulated from the aluminum cylinder heads... the charge cooler ends up getting thermal load from the engine as well. Phenolic spacers to isolate thermal conductivity, heat shields to reflect ambient heat soak, under-hood airflow management, as well as surface area and thermal mass of the charge cooling system itself all come into play.
But it does shorten the run, and reduce the volume of the charged air intake system, and reduce the compressor side of the lag factor of building pressure from the turbo compressor through large diameter, high-flow tubes, low-restriction, high surface area intercoolers, more tubes, past the blow-off valve, through the throttle-body, and into the throttled intake manifolds, cylinder heads, through the valve lift opening and filling the cylinders.
The more volume that is, the longer time it takes, and the more chances that something has changed during that time... It magnifies the issues with both smaller, low-lag turbos, which have a harder time building that intake pressure quickly, or larger turbos that introduce more turbine-side lag while exhaust pressures rise, and the turbo's heavier spinning assembly accelerates slightly slower, due to inertia. Every time the blow-off valve releases a large volume of air from the intake tract, that pressure has to build again.
There are advantages to making a short, high flow, but lower total volume intake tract work with a liquid charge-cooling system, if done right, especially if lag is to be avoided to more closely resemble an atmospheric engine's quick, direct responsiveness.