It's perfectly reasonable to employ a limping strategy at ~20bbs. It sounds like your friend has an antiquated line of thinking if that is a direct quote. When creating a limping strategy from ~20bbs you tend to want to do it more often (in terms of overall range %) from early or late positions. Ie. UTG/1 and BTN/SB, but will have it at some % from every position.
Is there merit to always raising? Yes. Is there merit to creating a strategy that involves limps? Yes. I'd say your strategy of adjusting to your opponents and capitalizing on how they will react can net the most EV. Though, we can't always assume what people will do and sometimes our assumptions can be very wrong. It's important for you to figure out your ranges. FWIW this sounds like a good spot to limp (position, your reasons about the player, etc.) and AQs is a good hand class to do it with.
For example here's a 20BB CO chart from DTO:
As you can see their solution has a 12.2% limp range. I'm not saying I recommend trying to implement a 12.2% limp strategy at 20BB on the CO, but it speaks to the advice you got.