The radioactivity accidents that are picked up by environmental sensors tend to be very large-scale. The little, nasty ones, like the mining source that was lost in Australia, tend to disappear.
Minuscule tritium emissions are detected around nuclear power plants all the time, far below the level at which anybody should be concerned.
There was even a case where alarms were sounded when a power plant worker was found to be radioactive due to radon in his home, which triggered detectors at work. The general chemical industry doesn't operate with anything even remotely close to this degree of care.
That's a gas, making it the easiest possible case. I think a more plausible threat are the decaying temporary storage containers that a lot of low-grade waste is sitting in because nobody can find a permanent location for it. (Yes, that's largely due to political reasons, but political reasons are real!)
A gas is quite difficult to detect because it diffuses. This also makes them typically less dangerous because diffusion also means dilution. Levels do matter. Tritium is also pretty low in terms of radioactivity. The radiation cannot penetrate the skin except in very high levels (you can find keychains, watches, gun sights, etc with small bits of tritium and phosphorous to create long term glow objects). Ingestion and inhalation are more serious since your internal organs are more susceptible (see weighted dosage). The real cool thing is that we can measure radiation with high precision and in real time, so we can detect these dangers. Mostly because these devices are cheap.
In addition to the requirement of a more active approach needed to detect ground/water contaminants there are also a larger variety of pollutants that are harmful. Many of these need specific tests, which can consume your samples. Of course we can do pretty good guesstimates for what we should look for, but we do need to recognize that the process is both more fuzzy and more involved. We can grow these projects by making them cheaper, but that's a tall order (it is happening though).
Edit: I do want to note that most radiation detection devices do not distinguish between types of radiation. These differences do matter in danger levels. This can add complications the above but there is a decent signal that is still useful. But as with everything, some expertise and domain knowledge is quite important.
From what I understand, in the case of the radon-contaminated nuclear worker, what they actually detected were the "radon daughters", the decay products of radon. Radon itself doesn't stick around very long, but when it decays it produces an atomic dust of polonium, lead, etc. That dust is what tripped the alarms.
Anything like that radioactive source from Australia would set of tons of alarms in a nuclear power plant. You wouldn't get it out the door. Incidents like that missing radioactive source in Australia happen where there are far fewer safeguards than at a nuclear power plant. Those sources generally go missing from abandoned medical equipment, food irradiation facilities, and that sort of thing. You'd be hard pressed to smuggle (let alone accidentally convey) something like a spent nuclear fuel pellet out of a power plant.
