The FCC is already on a pretty optimal answer in this case, it's physics you'd want to fix. Completely jamming all cell signals disrupts the public and private airspace well beyond the school.
It's difficult to explain accurately without making one into an RF expert but you've essentially said the radio equivalent of "my headphones aren't really useful past 10 feet" and concluding any sound so loud as to make the headphone use unintelligible across an entire school would also cease to exist a few feet away.
If one absolutely had to get rid of usable signals originating from outside the school a much more practical way would be to add RF shielding to the exterior elements. After all, if you want to get rid of a signal then subtracting it even intuitively feels a lot more reasonable than trying to add so much other signal you don't notice the original anymore.
Another way to jam would be to spoof being a tower. The cell phones would uselessly connect to the school tower. Since phones transmit their location, the spoofer could "geofence".
Or even ask the local cell phone tower people to geofence the school.
Spoofing is, by definition, not jamming but in terms of the goal if you combined a multi carrier spoofer with a DAS you could theoretically get things down to ~a million $ per high school on average and not horrifically leak signal as if you just stood up a spoofer and centrally blasted it out. It's also still carrier and technology generation specific though. I.e. if a new carrier/MVNO appears or an existing carrier rolls out new frequency bands then you need to replace the system. If it's a new technology (e.g. 5G) then you need to replace the system. If things just change in general (new MVNO on existing bands), you at least need to maintain the system accordingly (but might not have to replace it). You'd also need to convince the majority of people this is something all public companies should be mandated to support, which is a separate issue from convincing the majority of people this general approach should be taken, or you'll never get much compliance. All of that accounted for, the solution still brushes against needing to fix physics a bit in that lowers SNR in a many times larger area than the accepted dead zone - though not as badly as a jammer system.
Geofencing with cooperation of the local carriers is technically possible (with big downsides) in that if a phone is connected to 3+ towers you can decently get the position tower side to within several hundred meters (or worse) in an urban area. That means to reliably geofence a building you're still going to screw up anyone in the surrounding city blocks trying to drive by, use their phones at home/work, and so on regularly. It also still impacts noise (therefore connection quality), though not as bad, over a much large area than is excluded. You could also theoretically mandate that all devices have device side geofencing which uses the extra data from phone side (more than just tower connection data, BSSID or other local signal data) were you convince people to mandate public companies on this one as well (and even then there will be "old" devices for many years a la retiring 3G lest you want to convince everyone to throw their current devices away on top of the extra forced hardware monitoring ability). The cost for this would largely escape the schools by becoming a cost to carriers/hardware manufacturers.
Or you could, again, just build schools with RF shielding if this is really the path you want to go down. It'll probably have a higher up front cost but it'd achieve the goal for every high speed cellular service current and future without requiring any external mandates or external interference. It's still a bit silly compared to other funding requests schools typically have but it would solve the problem if the public prioritized it.
Many schools are set back a ways from other buildings.
I would just set up a jammer, and adjust its power so the nearest buildings could still use their phones, while the school phones would have problems connecting.
