Thanks for the comments. Yeah, I had already considered the mutliplayer-only aspect as problematic and thought of adding an AI months ago, but just haven't got round to it yet because AIs are pretty difficult (even dumb ones). I actually looked into reinforcement machine learning as a way of building a decent AI but it's very complex and difficult and basically requires using Python for everything, which I'm not that familiar with either. I'd love to get help from someone with AI building experience but I'm guessing those people don't tend to volunteer their time and it'd probably take a large salary to hire them :-)
Viewing post in No-one cares about my game
I noticed that for some board games, simple Monte Carlo Tree Search is surprisingly powerful. Very shortly (and badly) summarized: just simulate a lot of randomly played games starting from the current position, and pick the move that led to the highest percentage of wins.
Depending on how the game logic is implemented (with clear model/view separation), programming such an AI might be doable, though it works best for finite games, and your game definitely looks more complex than the simple games I tried it for.
Also, doing such a brute force approach might take a lot of CPU power, so doing it server side may become expensive. Anyway, it's just an idea that you might find useful...
Yeah given the tree of possible moves with my game that would turn into a virtually infinite number of games very quickly. Some kind of directed A* search is probably needed, and that is gonna be tough. Either that or hand-craft some AI behaviours that make some sense, but probably aren't all that hard to beat. As I said, ideally I'd have some kind of reinforcement ML expertise on my team and the AI could be trained to make good moves through the whole reinforcement neural net magic. :-)