Information architecture is how a site's pages are grouped, named and nested. Google's SEO Starter Guide says that organising a site logically can help search engines and users understand how pages relate to the rest of the site. It is also candid about limits: it advises against dropping everything to reorganise, because search engines will likely understand pages as they are now, and says the suggestions are most helpful over the long term, especially on larger sites.
The guide makes three concrete points. Descriptive URLs help, because parts of the URL can appear in results as breadcrumbs and help people judge a result. Grouping topically similar pages in directories can matter once a site has more than a few thousand URLs, because Google can learn how often the content in each directory changes and crawl directories at different frequencies. And each piece of content should be reachable through one URL, which connects architecture to duplicate content and canonicalisation.
Google's page on URL structure best practices adds requirements for a crawlable structure, such as not using URL fragments to change page content. It warns that URLs failing its criteria will likely be crawled inefficiently.
The strategic question is what the structure should be for: which sections the business needs, which searches each section should serve, and how the site will grow without becoming tangled. The technical side of architecture, including faceted navigation and large-site analysis, is covered on the architecture page under Technical SEO.