Sensemaking in Novel Environments: How Human Cognition Can Inform Artificial Agents