In Mercury, we do a lot of work with startups, where functionality constantly changes, and the code is rewritten from sprint to sprint. If you write an automated test, and then the team decides to make changes to the software, your test may quickly become obsolete, which means that instead