A lot of the language we use to describe evolution by natural selection is borrowed from the language of design, but natural selection doesn’t work the same way design does, so the metaphors are not very apt. Sometimes confusion results. Examples are “adaptation” and “purpose”. In design, a purpose pre-exists, and materials are adapted to it. In evolution, a change (mutation) occurs, and a use may or may not be found for it. If the change has utility, it may be by refining some pre-existing ability to exploit the environment, which looks a little bit like design (but isn’t). Or the utility of the change may only be relative to exploiting some previously unavailable (or less available) resource in the environment. That looks a lot less like design, and a lot more like seizing an opportunity.
For instance, an organism with splayed toes might occasionally acquire, as a genetic aberration, webbing between the toes. In a dry environment, this alteration might be dysfunctional (“non-adaptive”) and die out, or it might be neutral, and occur from time-to-time. To an organism dwelling on riverbanks, though, it might open new frontiers for (better) exploitation. If there is relatively little competition for the new, riverine resources, compared to the already well-exploited resources on dry land, the mutation may spread and become fixed, other mutations (changes in leg shape, body fat, feather type, bill type) which previously would have been useless or harmful now become beneficial, and may, if they occur, also become fixed. Speciation may occur.
This chance-and-opportunity model is not much like design. It’s more like exploration. In this view, ecological “niches” are less like problems waiting to be solved than like frontiers that some mutation may make available for exploitation.