They recognised the issue right away. Firstly, the bubbler pool was emptied of water (there was fear that if the core reached the water, it would cause a steam explosion). They started using liquid nitrogen to freeze the ground beneath the reactor, then scrapped that idea and filled the bubbler pool with concrete instead.
They started digging a tunnel in May 1986 [1] (7 days after the accident), with a void below the reactor. They planned to fit a heat-exchanger into the void. It's not clear whether it was ever installed, but the consensus is that if it was installed, it was never used, and the core didn't reach the void [2].
