Your specific example is a rather tame case. Ideally, you would need to incubate the sample in multiple batches to catch the bacteria in the sample that may not grow at a certain percentage of atmospheric oxygen. Testing an entire phage library against it can be automated and I've seen it done, but the machinery to do so is still what qualifies as a big investment for the companies I know of that work on this.
But if the library fails, and you need to find and mass-produce a new phage, the pipeline starts over again. That is what I was referring to in my previous comment. I have worked with certain gut bacteria that can be infected by phages, but those phages replicate in the tenfold range rather than the hundredfold range you see with bacteriophages like for E. coli. Growing such bacteria at bioreactor scale is challenging enough before you add in a phage that, by its nature, does not replicate at the scale you need. And post-reaction, filtering out the phage into something safe like saline can lead to more loss of phage, which I have also seen happen.
But if the library fails, and you need to find and mass-produce a new phage, the pipeline starts over again. That is what I was referring to in my previous comment. I have worked with certain gut bacteria that can be infected by phages, but those phages replicate in the tenfold range rather than the hundredfold range you see with bacteriophages like for E. coli. Growing such bacteria at bioreactor scale is challenging enough before you add in a phage that, by its nature, does not replicate at the scale you need. And post-reaction, filtering out the phage into something safe like saline can lead to more loss of phage, which I have also seen happen.