I wanted to wait with my feedback because i knew that the lat beta was seriously missing on important elements such as AI, therefore i waited for the second beta. It also gave me hope to the see that the developers got rid of certain gameplay mechanics that were just broken, instead of just 'fixing' them. I am speaking of the caravan system of course. Anyways the beta is still very rough and there are certain gameplay mechanics that will keep the game a balance mess until they are seriously reworked. I will come to those after a short list of bugs and small suggestion.

Bugs:

•  In my games the borders of wildlands do not go away even if i defeat all monsters including the boss monster . This happened to the fissure wildland and the old empire city wildland, i did not try other ones.
• Tireless march will stack up if you have multiple champions in one army, for example if you have an army with 4 champions ALL units will get 4 moves which also has a big impact in the combat map (such as making enemy archers useless).
• I can build forest and normal drake lairs with the empire although the tooltip says i can not. I can also build those units.
• The routine that draws the route that a unit will take when you give them a move order can use huge amount of cpu resources if the way is complicated. Just stand next to a long mountain range a give a unit a command to move to the other side of the mountains.
• There are still crashes after long periods of gaming (win 7 64bit + 4 GB) and it will nearly always crash after loading game from within a game.
• Forests still have there outlines even if i deactivate this option in the menu
• when zooming in from the clothmap, forests will sometimes be grey, like they are in some fog of war. In this case i have to zoom closer before they are colored again.
• Combat sometimes is still stuck and i have to use the autoplay (within tactical combat) feature to deal with that issue. It is a rare issue though.
• In the late game scrap golem armies will spawn in water

Small Suggestions:

1. The movement path dots should be colored differently for every turn a unit takes, for example they are green for the first turn, then red for the second, then blue for the third turn and so on. Currently you do not know how many turn a unit will take for reaching its destination.
2. There should be a game option where i have to click again to confirm a move command, i.e. the first click is to bring up the movement dots and the second click is to confirm the move order and only then the unit will move.
3. There should be some indicator so i can see how long it takes until a treaty expires.
4. Please rework the fallen terrain, it is really hard to look at for longer gameplay sessions.
5. For the love of god please fix the loot tables. In my last game my sovereign found a titanic maul in combat, in the very early game (i did just finished teching to leather armor). A good weapon and armor drop can win you the game because there is nothing stopping you at all.
6. Please make better artistic indicators for land that can be settled. Currently i can only know if i select a pioneer and hover over the settle button. I would love to see which spots are good for city placement, by just looking at them. In civilization you could tell good city locations just by the terrain.

Issues with core Gameplay mechanics:

It seems to me that there needs to be major overhauls to some of the mechanics and issues listed below , because currently their combination makes no sense gameplay wise. They might be nice ideas on their own, but as soon as you combine them it leads to horrible gameplay.

 a) Champions and Taxes:

This combination of gameplay mechanics is the current game breaker for me. The basic issue is very simple to grasp:

 Champions are a viable alternative to troops, but cost no upkeep.

This leads to the fact that i can consistently run on no taxes, enjoy high production and research while being totally competitive in military terms, even ahead mostly. No matter how good the ai will be eventually it will always be just a little toy and no serious threat, because it makes the big mistake of building units, therefore it needs taxes, therefore it will be behind in research substantially. You also do not need any military technologies, besides leather working, until very late in the game because once champions have leather armor they can find their own (and much better) equipment as well as being a big revenue stream on their own (all the useless looted items you can sell). This means i can focus more on economy technologies, so any ai economy will be a joke after a short amount of time. Once my economy is ahead and a threating (although i never met one) ai declares war i can usually surpass their military technology level in a couple of turns because of my superior economy/research.

b ) Building maintenance and taxes: 

Again a very simple issue, most buildings that increase research and production cost maintenance which means they are worthless. For example at first i often made the mistake building workshops, lumbermills and research buildings in most cities until i realized i have more production and more research when NOT building them! The reason is simple: If i do not build them i don't have to pay any money thus can keep taxes low and therefore have more research and production than with those buildings. I usually end up with one city that has all production buildings to pump out troops or wonders just in case i need them. The rest of my cities have only buildings that cost no maintenance, for example merchants, inns, pubs, farmers markets (a very important building to spam) ...

This could be fixed by giving maintenance to growth and unit training buildings, such as inns, pubs, command posts, etc. and remove it from production and research buildings. Currently in my games i make ~50-100 gildar a turn with no taxes and enjoy much higher production and research rates then my rivals although i have less cities.

c) Champion placement on the map:

To me the idea of placing champions on the map is very bad in itself. First you can simply kill the opposite alignment champions which not just denies them to the enemy (as if the ai would ever research the necessary technologies to get them), but you also get their weapons, which can be a real game changer. Secondly it also makes it very luck/ map dependent on what champions you can get. If you end up with a bunch of high level champions and fast tech to the recruitment technologies you can steamroll the map. I much prefer any other mechanic where champions can be bought off map,  e.g. in an inn and there would be technologies that limit the amount of heroes you can hire as well their starting level. If a champion has too many negative perks from loosing combat you (and the ai especially) should be able to retire them.Alternatively give neutral champions at least escort armies and don't let them drop their weapon when defeated.

d) The 'kill a champion to get their spells' spell: 

Sorry i forgot the name of the spell, i think it was steal soul.  Basically you sacrifice a champion to get their spells. In my current game i made a custom caster sovereign and gave him as much talents as possible and NO spells levels. Then i recruited 7+ heroes with spells leveled them a bit and sacrificed them and thus is got a powerful death mage, earth mage, fire mage, ... super melee mage steamrolling super sovereign. The basic problem is that you can not just get the basic (apprentice) spell level, but if you sacrifice the champions in the right order you can get to higher spell tiers and you end up with a sovereign that may be only be level 8 but has as many talents as a level 20+ sovereign and is a master of melee and magic.

e) Maul, especially on bears:

this only applies in the early game, but maul is much too powerful on bears. It just sucks to loose important champions or units in the early game to one bear that keeps on mauling. I would prefer if it would no just decrease in accuracy with each strike, but also in damage as a quick fix. currently it is very random and a bit of bad luck can have a big impact because you encounter units with maul mostly in the early game where your own units have too low a level, i.e. hit points to survive this bad luck. It is not that bears are too difficult to beat, it is just that with a bit of bad luck you can loose important units. In the late game a champion with maul is absolutely overpowered, but this is another story.

f) Tactical combat: 

Tactical combat has seen a big improvement since the last beta, just because maps now scale in size depending on the clashing armies size. However it is still not very tactical at all. I agree with the developers that the pace of combat needs to be fast, i.e. not taking longer than 5 minutes and the current beta does very well on this. On the other hand it could be made a lot more tactical by some small adjustments like:

1. All units have one default counterstrike (can be improved by equipment/talents)
2. Give certain equipment inherent tactical qualities, for example spears could give you first strike, i.e. their damage is calculated first even when counter-striking, shields could ignore one ranged attack per round, magical staffs have damage over time effects (fire) or reduce initiative and movement (frost), maces ignore a certain percentage of armor, ...
3. it would be nice to see terrain modifiers and especially line of sight for archers, i.e. you can't shoot units behind trees.certain units could also act this way, for example a drake could make friendly units behind him not target-able by archers, because he is so big.

edit:

g) Faction diversity: 

I am not necessarily speaking of the difference from faction to faction but the difference between empires and kingdoms. To me they play very much the same with some minor differences. Where are the unique tech trees? Why can kingdoms settle on fallen land and empires on lush green lands? Currently the differences are not really noticeable in gameplay, besides some diplomatic modifiers etc. the gameplay experience between the two should really be different in my opinion.