slowrotor,
As Chief R&D test pilot and also as a member of the design team for Comanche, I had a chance to really probe the question you ask (a really good one, and fundamental, yet ignored by most agencies in their pilot training literature). There are two kinds of load factor the designer must worry about:
1) the limiting loadfactor of the static structure - the airframe (static means it doesn't rotate). This is the big number that defines when the tail falls off, or when the engines might get out of their alignment. It is usually 3.5 G for a part 27 or 29 helo, and has almost nothing to do with what you can actually pull in flight. The design is subtantiated for the g level, and must be shown to have a 1.5 safety margin (3.5 times 1.5 = 5.25, the acutal g at which failure will occur). Often, crash load factors or landing gear loads are more important to the structure.
2) the actual thrust capability of the rotor and the rotating components, expressed in g's. Usually about 1.5 to 2.5 for most helos at sea level, this is what the rotor can actually pull, so this is the number you will see on a g meter if you do an aerial maneuver. The limit is rotor retreating blade stall, where the rotor can't produce any more thrust, or where control is nearly lost as the thrust is increased. Unlike an airplane, rotor stall can't be easily felt by the pilot. Do not believe the training guides, which give all kinds of advice on stall detection, based on 1953 helicopters. They are of little help for machines with advanced airfoils, where stall causes sharp change in the blade pitching moment, and therefore increases in the loading of the pitch links, and therefore can actually drive the controls around the cockpit (jack Stall). As you climb above sea level, the g falls off, and at the service ceiling, the g capability might be 1.1 g's.
Most helos are designed as transports, for carrying things and people, just as most cars are. Don't maneuver your helo agressively for kicks, any more than you wouuld take your car to Indy for some trials. the awesome surprise some folks showed on this forum when a Robbie blade was fould cracked after the guy was doing nothing more than "mustering" is an example.