I think the major change is scans are less certain and certainly not instant. I wrote a whole post with lots of details and formulae I came up with on the private forums, but tbh that wasn't the point. It boils down to the following:
Against a "top security" planet (i.e. spending a one off payment of 15k resources per asteroid you have to make your planet secure), you are waiting 2 ticks with a 66% chance of getting a succesful jumpgate scan back, or 1 tick with a 58% chance of getting a succesful scan back, or a 0% chance at 0 ticks (all would be 100% if the planet only had a 75% security rating).
You can send multiple agents, but I guess the agent-dying thing is a bit harsh for scans.
I actually like the idea more now though - how many agents are you willing to train up and send to make the chance of you getting a scan or covert operation more likely? "Scanners" will need more. Do you send 1 agent with a 66% chance of getting through? 2? 3? 5?
Also, blocked scans would probably show again, giving some indication of if your galaxy will get incoming.
Alliances will always be more organised - people will preorder jumpgates for their targets so they're ready, and unit scans can always be done before time (58% chance to instantly get a sucessful scan, 66% chance to get one done after 1 tick).
Espionage will just be a unification of scans and covert ops to some degree, housing them all together. They'll still be a bit different - I was considering rasing security levels for covert op missions in addition to making them take longer to complete.
scans would now have the randomness again though, and some would have a time delay. It would just generally remove the whole "launch scan recall" nature of the game.