I'm not really one for major nerfs...but you did touch upon what other schools lack: Efficiency. Let’s look at all spell schools and judge their worth from levels 15+.
Enchantment:
*Problem-You disabled the enemy, but typically need to kill them still. Nuking, AOE DoTs, melee DPS, ranged DPS, Instakills...something.
*Solution-Enchantment is in a very good place ATM.
Illusion-
*Problem-The school itself isn’t as much a problem as the number of spells in it. With only 7 spells (6 if you count a mass version of invisibility as the same spell) illusion doesn’t get its proper chance to shine.
*Solution- Add spells, adjusted from PnP versions.
Color spray
Silent/minor/major/Persistent/Permanent/Programmed/Project image: summon monster that doesn’t do anything but eat agro
Greater/Shadow conjuration
Greater/Shadow evocation
Simulacrum-Duplicate monster/ally
Weird- AOE Phantasmal killer
Conjuration-
*Problem: None really. This is a decent school.
Solution: none needed
Abjuration-
Problem: Mostly buff spells. Dispel magic is typically useless.
Solution: Lower the cost of Dispels and make it target spells that are beneficial to the enemy only. Likewise, when cast on an ally, make it only target harmful spells. Allow enemies to buff themselves in ways that make dispelling actually useful.
Evocation-
Problem: Becomes inefficient and slow around now. This was worked on by the spell pass, but towards higher levels, straight nuking becomes far too inefficient for trash and is only used on bosses because/when saving sp has no value on those fights.
Solution: Adjust the cost of Maximize and Empower on nuke spells; I would suggest making it a percent of the normal spell, instead of 15 and 25 SP before enhancements. I would aim for spells costing 3-4 times their normal cost without enhancements or the wizard capstone and ~2 times their normal cost with it. Encouraging casters to nuke and having it be an efficient method of killing enemies will lower the difference between high grade casters and high grade melees.
Transmutation-
Problem: this school is mostly buffs, with the occasional 'Save or X' spell.
Solution: none really. This is in a decent place; the buffs are great and the other spells are good enough to warrant being used.
Necromancy-
Problem: the instakills are debated over constantly.
Solution: improve other schools to be equally efficient as simply killing the target enemies, then bring the debuff potion of necromancy up to par. If nuking (which is accepted as the main alternative to instakills and Heavy CC of Necromancy and Enchantment respectively) could get as much done as a Finger of Death or Wail could as equal or lower costs, Finger and wail would both be smaller problems.
Spell damage VS cost is rather strange towards endgame on failed saves:
Meteor swarm costs 50 SP without metamagics and maybe...65 with full metamagic cost reductions. It does 6d3+18 fire damage and 2d3+6 bludgeoning damage: around 80 damage max
80*2.5(maximize and empower)=200 damage.
200*2(being generous with enhancements)=400 damage
So 400 AOE damage for 65 sp...roughly 6 damage per SP, per enemy. Assuming 10 enemies, that is 60 damage per spell point.
Wail, on the other hand, will do as much damage as they have HP, up to 1000 damage on enemies with the new Epic ward. On 3 enemies with 1000 HP each you just did 3000 damage for 50 sp. 60 damage per spell point. Yes, using Wail inefficiently was just as efficient as using Meteor Swarm efficiently.
Change that and you won’t need to Nerf Insta kills. People currently build for speed and efficiency. Instakills are currently both; if instakills were fastest but other avenues were more efficient, more casters, especially the ones that haven’t yet reached 40+ DCs, would nuke instead of trying to Insta kill.