Crawling is a pull model. A search engine returns on its own schedule, which means a price change or a new article can sit unnoticed for days on a site that is not crawled often.
IndexNow inverts that for the engines that support it. When a URL changes, the site sends a single HTTP request naming the changed URLs. Ownership is proved by hosting a text file containing a key at the root of the domain, and the same key travels with each submission.
Microsoft Bing, Yandex, Seznam, and Naver participate, and submitting to one endpoint shares the notification with the others. Google has not adopted the protocol, so Google discovery still depends on crawling and sitemaps.
Implementation is small. Most enterprise CMS platforms and CDNs offer it as a toggle, and a custom stack needs a publish hook that posts the URL list. The two rules that matter are to submit only URLs that genuinely changed and to keep volume reasonable, because repeated submissions of unchanged pages get throttled and eventually ignored.