The hero AI is very, very basic right now. He looks at neighbouring rooms only, and if there's one with treasure in it he goes for that. Otherwise he picks a random one! (this will be improved later).
The next gameplay improvement I'm thinking about is to improve what you do each turn (choose cards to play). I'm thinking of giving you more choice each time, but with consequences. So for example there could be a monsters & traps deck, and an items & events deck. You can choose which deck to draw a card from, but you will want to balance the good & the bad slightly. I'm thinking of either a 'danger level' for the dungeon that goes up and up, but playing monster/trap cards lowers it slightly (so you want to keep it at bay), or perhaps link it to the hero eg an 'interest' bar where if you never play any monsters he decides the dungeon isn't good enough and leaves.
I also have a lot of ideas about improving individual cards to be more interesting. For example a room that is a monster lair, or treasure trove. Special boss monsters. Etc
These are some great ideas on how to improve the game, and you're definitely on the right track. The problem right now is that there isn't a lot of tension - as you said it, not a lot of 'interesting decisions'. The ideas you have here are great for provoking such decisions.
A 'danger level' is much more visceral and entertaining than an 'interest meter', so I'd recommend that. Perhaps, if you don't play a 'bad' card for a while, the game begins drawing one random card from the monster/trap deck and placing it
for before you draw your cards. Further, as the danger level rises, the cards drawn from that deck could increase in danger, so you're given an interesting decision: do you place the monsters you're drawing yourself? Or do you push your luck and see what the deck gives you?
I also think there should be more risk - and reward - behind placing creatures down. Perhaps some monster cards could have a 'treasure' or 'encounter' card attached to them; "Kill me and you get this fancy 1-time use spell!" That would encourage people to take more risks. Further, instead of a simple 'monsters killed' meter, why not something like "Glory"? Then you can actively rewarded for trying to take down the biggest and baddest monsters.
Speaking of risks... many board games have random mechanics, and there's a reason for that. As it is, you can usually calculate your turn in advance, and know
exactly what's going to happen. Logically, this seems more satisfying to the user - everything's in your hands, you won't lose out of nowhere, and you don't suffer those "feel-bad" moments. But in fact, randomness is essential to much of the fun of many games, and I think this is one such game. When you see a card that has a
chance of hurting you, you go - "Do I risk it? Do I try to get the reward? I should be fine..." Then you eagerly watch in anticipation - maybe even a bit of fear - to see what the result will be! As it is, no such tension happens with this game. The drawing of the cards is pretty much the only luck factor here, but the amount of control the player has over card placement negates such tension (another argument in favor of the 'wild placement of a bad thing every turn' idea). Some cards along the lines of "This demon offers a tantalizing offer... there's a chance of it backfiring on you, though. Do you take it?" or "The dragon has a 1/6 chance of casting a super-fireball when it attacks!".
TL;DR: While leaving everything in the player's control seems like a good idea, often it makes a game less exciting and removes the tension. I'd suggest exploring ways to craft moments of suspense, tense gameplay, and perilous decisions. As it is, you have too much information available, and exert too much control over the boardstate, to have any such tension.
This game is a wondrously beautiful idea, and I am eagerly looking forward to future builds, and sharing it with my friends. Best of luck!