Here's my experience with Cisco Works 2000(CW2K).
At a former company, we bought a BUNCH of new Cisco switches. Our sales rep threw in a free copy of Cisco Works 2000 on the order. My boss said, "make it work." So I got to implement it.
Now, {disclaimer} this was 3+ years ago. So the current product may be better ( I would hope so

).{/disclaimer}
In short, it was painful and not very beneficial.
Our environment included ~20 Cisco routers and ~50 Cisco 6500-series switches. I had been using expect(1) scripts to push all my configuration changes.
CW2K had a database back end that was temperamental. It took a lot of my time to keep it up and running.
The thing that looked to be of help was that it would report when any device's configuration changed. You could also schedule a push of a configuration change to happen when you're at home sipping a cold one. The process of setting up a config change was not very intuitive and I never got it to do what I think it should have been doing.
I still prefer expect(1) scripts to do my config changes.
I've since written a perl script that will check the "NVRAM config last updated" string, compares it with a stored value for that device, and if different, kicks off a scarfconf.pl process to download the newest config version. (scarfconf.pl comes with the current distribution of MRTG under the contrib/cisco_tftp directory.)
The only other feature of CW2K that I didn't have was the Response Time Reporting (RTR now called SAA) tool. This provided the ability to see and report the delay between RTR-enabled devices. But this was before Tobi had smokeping fully developed. (see
In the end, I convinced my boss that we weren't getting much return on the amount of effort we had to invest to keep it running. It became shelf-ware.
HTH,
Patrick
Patrick Bartkus, CCNP, CNX, SCM, RHCT Sr. Network Engineer
GA Dept of Labor IT Network Services
If truth were not absolute, how could there be justice?