URL migration is one of those tasks that keeps people up at night. Hundreds of subpages, redirects, hreflang, sitemaps, the risk of losing your Google rankings - in the classic model it is a project measured in weeks and a serious agency budget.

We recently moved two brands to a clean /pl + /en structure at once (several hundred URLs), in a single afternoon, with no downtime and without a single broken link. The brands are VineyardElf and WineryElf - technology platforms for vineyards and wineries from our own stable, running on the very same engine that powers the new Siteor. And it was that engine - a CMS built around an API - paired with AI that turned the tedious, error-prone part of the work into a non-issue.

Here is how it looks in practice and why this combination changes the rules of the game.

Why /pl and /en (and not separate domains)

For a single brand with multiple languages, the best structure is subdirectories: yourdomain.com/pl/... and yourdomain.com/en/.... That is how Stripe, Apple and IKEA do it. The reason is simple: all your SEO authority stays on one domain, and Google gets a clear signal about which content is in which language. The alternatives (separate domains or subdomains) dilute the domain's strength and cost more to maintain.

The catch is that switching to such a structure touches literally every single subpage. And that is where the real work begins.

What you have to handle (and what usually goes wrong)

A proper migration is not just changing addresses. It is a bundle of interconnected pieces, each of which can quietly break:

1. Changing the paths of every subpage to /pl/... and /en/...
2. 301 redirects from old addresses to new ones - with no chains (an old address must not hop through three stops)
3. hreflang and canonical pointing to the new, correct addresses - not the pre-migration ones
4. A sitemap with the full set of new URLs and submitting it to Google Search Console
5. Redirecting the homepage to /pl or /en depending on the browser language
6. A language switcher that leads to the equivalent page, not the homepage
7. Links inside the content - the ones embedded in articles and subpages need updating too
8. Analytics - checking that Google Tag Manager has no rules pinned to old paths

Do any one of these by hand across several hundred pages and you are guaranteed a typo, a forgotten link, or a canonical pointing to an address that no longer exists. And every such error is a real loss in Google.

How Siteor does it together with AI

The secret is that Siteor is API-first: every page, every layout, every article is accessible and editable through an API. That means AI can treat the whole site as data - read the state, calculate the changes, apply them in bulk and verify the result immediately. Instead of clicking through the panel hundreds of times, AI works on the entire structure at once.

In practice it went like this:

Discovery. AI pulled the full inventory of pages through the API, identified which were Polish and which English, and computed the exact list of changes. Surprises surfaced right away - for example, some older subpages had old addresses hardcoded in their hreflang data. Without that scan they would only have come to light in Google, weeks later.

Execution. Changing paths to /pl and /en, remapping redirects so that every old address leads in a single hop to its final new one (zero chains), updating hreflang and in-content links, generating a complete sitemap, setting up the homepage redirect by browser language. All of it consistently, with a backup before every change.

Verification. This is the part a human usually skips out of fatigue - and AI does it meticulously. After the migration it swept the entire site and checked:

• whether canonical and og:url point to the new addresses (zero leftovers from the old ones)
• whether there are any redirect chains
• whether every URL in the sitemap returns 200 (not a redirect or a 404)
• whether a single broken internal link remains in the content
• whether the language switcher leads to the correct equivalent
• whether Google Tag Manager has any rules pinned to old paths (it did not - analytics tracks by address automatically, so the migration left it untouched)

The audit came back clean. And wherever something did stick out - for instance individual subpages still rendering in Polish because their English version was not finished - AI simply translated them, preserving the entire layout, proper names and industry terminology.

The result

Two brands fully moved to /pl + /en. Several hundred addresses, a complete set of 301 redirects, correct hreflang, sitemaps submitted to Google, a language switcher that leads to the right equivalent page. No downtime, no lost links, in one afternoon instead of weeks.

Takeaway

URL migration does not have to be a project you dread. When the CMS is built around an API - like Siteor - and AI takes on the dull, risky work, you get something the classic workflow cannot deliver: scale and thoroughness at the same time. AI does not get tired at the hundredth subpage, does not forget a redirect and does not skip verification. And you get a result that would normally require a team and an agency budget.

If you run a multilingual site and you are thinking about tidying up your addresses - this is exactly the kind of task where Siteor paired with AI shows its teeth.