Crawl budget
Crawl budget is how many pages a search engine can and wants to crawl on your site in a given time. It matters most for very large sites.
Crawl budget is how many of your pages a search engine's crawler can and wants to read in a given stretch of time. Google describes it as the set of URLs it can and wants to crawl, made of two parts: how much crawling your server can take, and how much Google wants to come back for your pages.
For most sites it's not something to worry about. Google's own guide to crawl budget is written for sites with roughly a million or more pages, or tens of thousands of pages that change every day. If your site is smaller, new pages usually get crawled without any help.
What wastes it
The usual problem isn't having too many good pages. It's having lots of addresses that show the same thing, or nothing at all. Filters and sort orders that each create a new URL, long redirect chains, and pages that return errors all take crawler visits that could have gone to the pages you care about. A slow or failing server makes it worse, because crawlers back off when a site struggles.
For example, Harrow & Pine has 400 products, but every colour, size and sort option makes its own address, so a crawler sees over 20,000 URLs. Most of its visits go on near-identical lists, and a new product page can wait days to be read.
Crawl budget and AI bots
AI companies run their own crawlers, and each one decides how much of your site to read. A page an AI search bot hasn't read recently is one its platform can't quote from with any freshness. The same fixes help every crawler: fewer duplicate addresses, no error pages linked from your site, and a server that answers quickly.
In AnswerOrbit
Connect your server logs and the Bot visits tab in Website shows which AI bots visited, which pages they read and which requests failed. Recommendations can suggest fixing pages that return an error to bots. See Website and AI traffic and Recommendations.