Yes. The important words are "some" and "may".

The manufacturers needs to do fishing - trying to present their products in a way that gains them the maximum number of customers (or at least the maximum number of sold units). That means that they are not always 100% truthful. If a processor "may" be used for a specific application, they will not inform a customer that a competitors product is actually much better suited for the task.

That's why it is so important to be critical, and to read the datasheets and not just application notes or sales material. Reading the datasheets gives us needed information to try and - at least in our heads - map the actual processor to the problem we have in front of us. But the task is hard, since we need to abstract away a lot of pesky details while at the same time not miss any show stoppers. We need to detect when a GPIO pin also have two alternative functions - functions that we need at the same time. We need to detect when the amount of code will be too large, or too slow. We need to detect if a peripherial is missing dual-buffering, or if activation of feature x will also push us above the required power budget.

The task is hard, but to be competitive, we must be able to browse 20 different chips, and then shortlist 2-4 candidates for closer examination. Even if one of our current processors can solve a problem with more or less tweaking, it may cost us too much to use it. Fighting lack of memory may delay shipments with weeks or months. Lack of I/O pins may cost us external multiplexer chips. Lack of configurability may cost us x mAh/day because we can't turn off a main oscillator while keeing an RTC running. Lack of availability may require the first redesign six months after initial shipments, or months of delays caused by lead times. Requiring 2.6V supply voltages just to manage the power budget may result in costly and space-consuming level shifters instead of having a different chip that manages the power consumption at 3.3V or 5V.

Neither of these things will be mentioned in the manufacturers sales information. The chip may be recommended for automotive use. But only when you buy the version with the extra letter in the name and the $1 higher price - but the price they initially show you is the "normal" consumer-grade chip. And they may tell that xxx only have 4 weeks lead times, but forgetting to mention that the xxxVX (the one with automotive classification) has 30 weeks lead time...

In the end, you'd like some good relations with some brand-neutral application engineers with access to prerelease information for new chips and access to the order systems.

In the end, it's hell on earth to do a good analysis, since no manufacturer or sales organisation is too interested in informing about weak spots.