Are we talking 100 years out? Because you'll still need engineers to decide where to put a building, what it should look like based on financial restrictions (a robot reading your bank account to decide how much building you can afford?) and future demand projections.

The robots aren't doing the financial decision making in the next 35 to 50 years. Assisted analysis (software), sure.

A robot isn't going to understand civil codes when it comes to laying down power lines (humans barely understand the civil codes). A human would be required to follow such guidelines and then could perhaps instruct a robot toward the actual labor.

Maybe one day in a very far off future civil laws will be understood by robots, but I'm betting that's not going to be very high on the list of things to get around to automating. It would also likely require extremely advanced AI, because the robot might have to negotiate civil code conflicts based on decisions made (and this just seems far beyond the next few decades). Who takes responsibility for a robot failing to navigate zoning laws properly?

Were not that far from mostly automated architectural design. This is from an article(in 2005) about an automated system to design custom metal buildings[1]:

"The automated engineering done with the EDS, which is still being extended, is close to 98% complete for the simplest buildings and 50-60% complete for the most complex buildings. The tasks that are remaining are not automated, as they are either highly custom, or occur so seldom that the added knowledge capture for these few occurrences does not justify the automation cost."

There's also a similar tool for design of private construction[2].I think the software even understand civil codes.

[1]http://www.aecbytes.com/buildingthefuture/2005/RCCstudy.html [2]http://www.aecbytes.com/feature/2005/Norway_prefab.html