This is one of them instances where the idea sounds simple but really is not. There are many tasks you need to resolve, some already talked about.
1) How to model the office.
This needs to be done so that it is easy for you to model but the file is small. It has to be in a format that your path finding algorithm can understand. This means as you develop the path finding you will be making changes to the data.
2) Location finding / input. Done by hardware (gps will not cut it) or user says where they are.
Could be done with NFC and tags placed around in the office, not an ideal solution, the tags have to be visible. The user could input their location but they may be lost and not know where they are.
3) How to present the data, you could write a magical 3D app showing a mission impossible building layout. That will be six months work in it self! Remember, your map is in 3D, road maps tend to be 2D. (with some fly overs...)
4) Progress tracking. The app really needs to know how you are doing so if you go off the route it'll tell you to go back. This again will be hard to achieve indoors. Progress tracking is one area that PND (Personal Navigation Device) manufactures spend time on as it really effects user confidence in the software.
Having worked at TomTom all I can say you could be at this for years.
As this is a collage project could I suggest a simpler project?
Do it with 2D map, simple grid based map a bit like the classic mouse in a maze problem. Implement A* for the path finding. Once done you could then if you have time extend it for your original plan.
A book I have found very good in the past is "Programming Game AI by example". ISBN 1-55622-078-2. Chapter 8 is all about path finding and has a very nice part on A*.
Good luck. :)