science

It depends A LOT on your country and health system / coverage among many other things. But here's my two cents based on a country with public healthcare.
It is a common cathegory in neuropsicological evaluations when a person starts refering to cognitive difficulties that may or may not be quantifiable as of yet. We put this group under "cognitive complain", sometimes their evaluation reveals actual changes in cognitive performance, sometimes it doesn't. This group sometimes progresses throughout the years into dementia, but not always, since many factors can impair your cognitive habilities (stress, depression, anxiety, other conditions like MS, etc.).
The laboratory could benefit then from providing a bunch of hospitals that perform these evaluations with the tests to diagnose the profile they need for the medicine to work (as Ive seen done in my country with many conditions, including rarer types of cancer and rare genetical disseases). The lab benefits cause otherwise, no one would buy the drug and no coverage would approve it without proof its going to work. This works becuase anyone, money or not, can schedule an evaluation and these hospitals when they have cognitive complaints.

There´s also clinical trials and or other research projects usually in place in these kinds of hospitals, already studying and performing various tests on dementia patients. Elly lilly just has to find and support a PHD project that studies amyloid buildup and then everyone participating gets tested (and then referred somewhere for treatment with the drug). I dont wanna reveal too much of my life on the internet but Ive seen something awfully similar to this first hand in my own country, a bunch of families got genetical testing for dementia for free, then everyone positive to certain genes got referred for treatment.

As research progresses, some other laboratory may come up with a drug that works for a different type of dementia profile, and then it gets easier because they often strike up deals between the two laboratories. They both eat up the cost of testing, since one of them is likely to benefit from it. It is already done this way for genetical testing for rare disseases (again, we are talking a country where the patient does not need to pay or just pays their monthly fee for private coverage but no extra for these things cause they are covered by law).

Then all they need is to get their medication approved and included in the lists of medicines that are covered by law. There's health coverages specifically for older people that often get deals with the laboratories of medications their patients are more likely to need. you get the ball rolling, probably a patient support program that can help with adherence and provide guidance to the patients and their families as to which documents to submit and where to get their medication covered.

What I mean is, its going to be different in different places, but there are already plenty of medications that require prior testing (sometimes really expensive testing) so the pipeline is already in place. I know nothing of amyloid buildup, but if there's any chance its genetic, then the easier it will be (cause then you take current dementia patients and test their families).

I don’t agree that this is the real deal

What I meant is that we finally have a disease-modifying drug for humans. It opens a lot of possibilities. I do agree though that benefits for the patient are limited for the time being (unless we can prevent Alzheimer's for pre-symptomatic symptoms, but that remains to be proven).

Is amyloid testing difficult/expensive?