When Stantec deployed AI agents on a 190,000-square-foot health care project, this question stopped being theoretical: Where does AI execution end and architect authorship begin? This session—co-presented by Stantec's senior BIM administrator and SWAPP's CEO—will walk through the actual decisions made on that project. We’ll look at which documentation tasks the AI was allowed to execute, which were held back for human review, and what happened when AI agents moved a fixture during an ADA compliance check. You’ll leave with a workflow categorization framework that separates judgment-dependent work from execution-dependent work, applied live to 20 common tasks, plus the standards-governance playbook used to keep firm authorship intact under AI execution.
