Computer coach at chess drills

Could the computer coach at chess drills be set to behave more like a human?

Sometimes, as in the Lucena position drill, the computer will just "abandon" the defense and hand over a piece. It just doesn't put up a fight like a human does.

For instance, I've been playing this Lucena drill (https://www.chess.com/drills/practice/rook-endgame-lucena-position)  and I don't often get to 'build the bridge" as is expected in the technique to this endgame. 

Some computer responses include:

1. Rd1+ Ke7 2. Rd4 Ke6 3. Kc7 Ke5 4. Rb4 Ra7 5. Kc6 Rxb7 6. Rxb7 (giving up the rook)

1. Rd1+ Ke7 2. Rd4 Ra3 3. Kc7 Rc3+ 4. Kb6 Rg3 5. b8=Q Rb3+ 6. Kc7 Rc3+ 7. Kb7 Rb3+ 8. Kc8 Rxb8+ 9. Kxb8

Wouldn't it be better if computer behaviour attended to a tablebase, instead of being let "free" to do what it considers suitable in the position?  It feels like the computer acesses it will lose anyways, and just gives up...  The consequence is that we don't get to practice the technique we are supposed to practice in drills such as the ones I mentioned above.

Of course, this doesn't happen when the drill is a defense drill, such as in the Philidor position. Then the computer keeps the pressure up.