There's no point running the tests if noone can see or interpret the results. My chosen approach is to write results out to a SQL database and provide access to all interested parties via a number of web pages.
This approach uses several technologies:
Automation doesn't just apply to the execution of test scripts. The results of the scripts are essential data and this data can be transformed into information via automation.
For instance, in a financial modelling tool, you could plot failures against functions, giving not only areas to concentrate on for debugging, but, if your models are life-like it suggests where you could concentrate your optimisation efforts.