The part will grow due to the change from pearlite and ferrite to martensite (quenching produces a martensitic structure in steels). The size change is affected by the alloy carbon content. Lower carbon steel expands less than high carbon. As stated above, you will not likely see much change in 8620.
I agree with this 100%, but since the OP is a newbie, he needs to know don't count on this in every situation.
I have made thousands of small (.125 to 1" OD ) 4340 parts for nearly 20 years that get quenched, then tempered to 30 HRC +/- 2. They are short, sold and cylindrical with tolerances ranging from + /-.0005 to +0/-.0005. They used to be done in oil when i first started, in a vacuum for the last 12 or 13 years. Both ways, and done by several different heat treaters, all of the parts have shrunk .0002" every time over hundreds of batches. So much so we count on it and shoot for .0002" over nominal.
Hollow or thin parts are a different story. Very unpredictable and varies with design and heat treater based on the process, furnace, support or fixturing, etc., as another poster advised.
Either way, if you have to be exact, either grind or machine post heat treat, or make a small sample batch, see what happens, plan accordingly.
I also can't stress the importance of finding a good heat treater, as was previously posted. They are not all equal and you may have different needs in different situations. I have a cheap guy that i use when it isn't precision size or hardness and has some margin for error. i have another when minimal distortion and close hardness is crucial. It took me years to find a cheap guy who wasn't terrible and a top notch guy who really is what he's cracked up to be, not just all hype and exaggerated promises.
Bottom line, if you have a few thousandths to play with, heat treat is rarely a big deal.
Sorry to get preachy and long winded, i do it a lot because i wish more people had been more detailed and offered more advice than i knew to ask for when i was new.
Sent from my KFGIWI using Tapatalk