In a domain environment, it is usually handled on the domain controller. DHCP hands out addresses, and the workstations are authorized to update their record in DNS.
DHCP will handle all the problems with IP conflicts, removing unused IPs, etc. It is also used to hand out other information such as default gateway and DNS servers. Otherwise, if you change one of those, you have to go to EVERY workstation and update their config.
Frankly, if they're paying you even minimum wage, it would quickly become more cost-effective for them to buy some variety of server that does nothing more than DHCP.
=P
Generally speaking, I tend to do a couple things...
If you have the luxury of using a 16-bit netmask (255.255.0.0) you can use the 3rd octet to signify a group of devices, and the 4th to designate the individual devices.
Example:
192.168.0.xxx for network infrastructure like switches and routers
192.168.1.xxx for Servers
192.168.2.xxx for Printers
192.168.10.xxx through 192.168.15.xxx for workstations
Alternately, you can just decide on particular ranges in a subnet for use:
192.168.0.1 through 192.168.0.10 for network infrastructure
192.168.0.11 through 192.168.0.29 for Printers
192.168.0.200 through 192.168.0.254 for Servers
192.168.0.30 through 192.168.0.199 for Workstations
Personally, I prefer using a combination of DHCP with the first method. DHCP hands out the workstation IPs, and then you statically assign the Servers, Switches/Routers, and Printer groups.