A network engineer configures policy-based routing on R1 to send traffic from the 10.1.1.0/24 subnet out a secondary ISP link. The route map uses 'set ip next-hop 203.0.113.2'. The primary ISP link is fully operational and the 203.0.113.2 next-hop is reachable via a directly connected interface. Users report that when the secondary ISP router (203.0.113.2) fails but its interface on R1 stays up, matching traffic is black-holed instead of falling back to normal routing. Which configuration change ensures PBR reverts to the routing table when the next-hop becomes unreachable?