There would be a substantial benefit to using a workstation card for professional use.
For starters, all GeForce Fermi and Kepler cards have their double precision floating point capabilities capped at 1/8th the performance of single precision floating point computations when the architecture is capable of performing them at 1/2. That's a four fold increase just for going to Quadro/Tesla if you use double precision.
Second, Quadro/Tesla have built in error correction. This is handy when running a simulation/render that can take hours to produce results. Tesla and Quadro chips are binned from higher quality chips and are statistically less error prone than GeForce.
Third, there are hand optimized professional driver tweaks for each CAD application that the developers pay good money to have included in the mainline driver
Fourth, the workstation cards are often overflowing with additional video memory that isn't available on consumer models.
