I'm not directly involved, but here's my understanding:
We use software by Zilliant. In simplified terms, it allows users to define regional markets and evaluate pricing on a product by product basis within each market. So, when a salesperson is selling a product in that market, there is a target price for the "average" price to expect in that market, but the system also provides a grid of prices around that target price, with the grid suggesting price "corrections" based on the nature of the job.
For instance, if it is a small job, the grid will suggest a higher price, because there probably won't be a long line of competitors chasing the same business. For bigger opportunities, the opposite is true. At the end of the day, the salesperson has the flexibility (within a specified range) to choose the price offered to the customer. However, the salesperson's commission is tied to the margin on the job, so s/he is incented to squeeze as much out of the customer as possible, while still earning the business.
Every quote, whether we win or lose the business, goes into the database to tweak (optimize) the regional pricing.
Hope that gives you some idea.