When you hire an agency to replace WordPress with an AI-built site, the AI is the least important thing to evaluate. What protects your traffic and your budget is the migration process around it: a crawl of the current site before the quote, a redirect map for every indexed URL, full ownership of the code and domain, a clear editing workflow after launch, and a defined monitoring window. WPBuildAI is one of these agencies, so read this with that in mind, but the checklist below is the one we would hand a friend choosing between us and anyone else. It covers what to ask, what good answers sound like, the red flags, how pricing should be structured, and when you should not hire anyone at all.
What does “replacing WordPress with AI” actually mean?
The phrase covers three very different services, and agencies rarely say which one they sell.
The first is an AI page builder with a migration bolted on: the agency generates a fresh site on a hosted builder, copies over some text, and points your domain at it. Fast and cheap, but the result lives on someone else’s platform, and the migration part is often thin.
The second is an AI-assisted rebuild into real code: the agency uses AI coding tools to write the new site as a framework project, typically React or a static site generator, stored in a repository you own. The AI makes the writing faster; engineers still design the structure, check the output and run the move.
The third is a content rewrite labelled as a rebuild: the same site, new copy generated by a language model. That is not a migration at all, and it carries its own risk, since Google’s guidance on AI-generated content rewards helpfulness, not the method, and mass rewriting can drop pages that ranked for their old wording.
Ask any agency which of the three it is quoting. The answer decides almost everything that follows, including who owns the result.
Which questions should you ask before you sign?
These are the questions that separate agencies who have run real migrations from agencies who build nice new sites. The right column is what a good answer sounds like.
| Question to ask | Why it matters | A good answer sounds like |
|---|---|---|
| Will you crawl the site before quoting? | Page count and URL patterns drive the real cost | ”Yes, we crawl first and the quote follows the crawl” |
| How do you handle every indexed URL? | Lost URLs are lost rankings | ”A redirect map, one row per URL, tested after launch” |
| Who owns the code, domain and hosting? | Ownership decides whether you can ever leave | ”You do, in your accounts, from day one” |
| How do we edit content after launch? | A site nobody can edit gets abandoned | A demo of the actual editing workflow |
| What do humans check after the AI writes code? | Speed without review breaks things quietly | Named checks: metadata, schema, links, accessibility |
| What happens in the weeks after launch? | Most migration damage shows up later | A monitoring window with Search Console checks |
| What is not included? | Scope gaps become change orders | A written exclusions list |
A vague answer to any single question is not disqualifying; agencies have different strengths. A vague answer to the redirect and ownership questions together is, because those two are where migrations lose money for years.
Why does a crawl-first process matter more than the AI?
Because the expensive failures in a platform move are almost never about how the new site looks. They are about URLs that existed on the old site and do not resolve on the new one.
Google’s site move documentation describes the process plainly: map old URLs to new ones, implement server-side redirects, update internal links and sitemaps, and monitor. None of that is AI work. It is inventory work, and it starts with knowing every URL the old site exposes, including the ones nobody remembers: old campaign pages, tag archives, attachment pages, PDF links that other sites point to.
We wrote about the method in what a crawl-first migration is, and about the artifact that comes out of it in how to build a redirect map. Ask the agency to show you a sample of either from a past project. A real one is a spreadsheet with hundreds or thousands of rows and a status column. If the agency has never produced one, it has never been accountable for one.
Who should own the code, the domain and the hosting?
You should own all three, outright, from the first day.
The domain is the easy one: it should stay registered in your name, with you holding the registrar login. Never let an agency transfer it into its own account for convenience. You can check the current registrant with ICANN’s lookup tool in a minute.
The code is where AI rebuilds differ most. If the site is real code, it should sit in a repository in an account you control, with the agency added as a collaborator rather than the owner; GitHub’s documentation on repositories covers how ownership and access work. If the site lives only inside a hosted builder, ask what export looks like and test it before signing. Some builders export cleanly, some export static files without the editing layer, some do not export at all.
Hosting should be in your name, or at least on an account the agency can hand over in an afternoon. The pattern to avoid is the agency that hosts on its own infrastructure, bills you monthly, and quietly becomes the only party able to run the site. A cheap rebuild that you can never leave is a subscription with a large exit fee.
How should pricing for an AI rebuild be structured?
AI tools have genuinely lowered the cost of building pages, and a good agency passes some of that through. What should not get cheaper is the migration work, because it is still a person checking URLs, metadata and redirects.
A healthy quote separates three things: the build, priced by scope after the crawl; the migration, priced by URL count and complexity; and post-launch care, priced as a defined window rather than an open-ended retainer. Fixed prices are fine and often better for you, as long as they are fixed after the crawl rather than before it. A fixed price quoted from a five-minute look at your homepage either includes a large margin for surprises or excludes the migration work you actually need.
Our breakdown of what a Lovable migration costs shows how those components add up for a typical small business site, and the same structure applies to any agency’s quote.
What are the red flags?
Most red flags are absences rather than bad answers. Watch for:
- A price before a crawl. The agency cannot know your scope.
- Guaranteed rankings. Nobody controls Google; an honest agency promises a process and monitoring, not positions.
- No mention of redirects until you raise them.
- “You will not need to edit anything.” You will, within a month.
- Code that only exists on the agency’s platform, with no export path.
- No plan for after launch, as if the job ends when the DNS changes.
- AI as the headline in every answer, with no description of what people verify.
- No accessibility mention. A rebuild is the cheapest moment to meet WCAG, and an agency that ignores it will hand you a site that needs fixing later.
| Red flag | What it usually means | What to ask next |
|---|---|---|
| Price before crawl | Migration work is missing or padded | ”What does the quote assume about page count?” |
| Ranking guarantee | Sales-led, not process-led | ”What exactly do you monitor, and for how long?” |
| No redirect plan | Builds sites, rarely migrates them | ”Show me a redirect map from a past project” |
| Agency-only hosting | Lock-in by design | ”How do I move the site without you?” |
| No editing demo | You will be dependent on them for changes | ”Show me how my team updates a page” |
When is hiring an agency the wrong call?
Two situations, and they point in opposite directions.
If your site is small, has little search traffic, and the content is a few pages you could rewrite in an afternoon, you may not need an agency at all. A hosted builder and a weekend can be enough, and the migration risk is small because there is little to lose. Check Search Console first: if the site gets a trickle of clicks from a handful of pages, a careful do-it-yourself move with redirects for those pages is a reasonable choice.
If WordPress is doing heavy lifting, complex WooCommerce rules, memberships, bookings, a multi-editor newsroom, then replacing it is a software rebuild rather than a website migration, and most “replace WordPress with AI” offers are not scoped for that. In that case the honest answer is often to stay and maintain better. We say so in audits; our post on when not to migrate lists the cases.
Between those two sits most of the market: brochure sites, service businesses, content sites and small shops where WordPress is mostly overhead. That is where hiring a migration-literate agency pays for itself, and where the checklist above does its work.
How do you compare two agency proposals side by side?
Lay the proposals next to each other and check that they price the same scope. The most common comparison mistake is choosing the cheaper proposal that quietly excludes the migration.
Check that both include: the URL inventory, the redirect implementation and testing, metadata and structured data carried over, the content editing setup, the analytics and Search Console reconnection, and a monitoring window. If one proposal includes all six and the other includes two, the price difference is the cost of the missing four, and you will pay it later in lost traffic or change orders.
Then ask each agency for one reference from a migration, not a new build. A client who moved an existing site with rankings can tell you whether the traffic held. A client who launched a brand new site cannot.
Key takeaways: hiring an agency to replace WordPress with AI
Judge the process, not the AI. Ask for a crawl before the quote, a redirect map for every indexed URL, ownership of code, domain and hosting in your own accounts, a demonstrated editing workflow, named human checks after the AI writes code, and a monitoring window after launch. Price should separate build, migration and aftercare, fixed after the crawl rather than before it. Treat ranking guarantees, missing redirect plans and agency-only hosting as red flags. Skip the agency when the site is tiny and low-traffic, and skip the migration when WordPress is running real application logic. For everything in between, WPBuildAI runs exactly this process, and the same questions are the right way to check us too.
Questions people ask before moving
What should I look for in an agency that replaces WordPress with AI?
Look for a crawl of your current site before any quote, a written redirect map for every indexed URL, full ownership of the code and the domain, a clear answer on how you edit content after launch, and a defined monitoring window once the new site is live. WPBuildAI works this way, and any agency worth hiring should show those five things in writing before you sign anything.
Does it matter that the new site is built with AI?
Less than the sales pitch suggests. AI tools speed up writing components and pages, which is why the work can cost less, but rankings survive or die on the migration process: URL mapping, redirects, metadata, structured data and testing. Ask what humans check after the AI writes the code. An agency that cannot answer that is selling speed without the verification that protects your traffic.
Who should own the code after an AI rebuild?
You should, completely. The repository should live in an account you control, the domain should stay registered to you, and hosting should be in your name or transferable in an afternoon. If the agency keeps the code on its own platform and licenses it back to you, the cheap rebuild becomes a subscription you cannot leave without paying for a second rebuild.
What are the red flags when hiring a WordPress replacement agency?
A fixed price quoted before anyone has crawled your site, promises of guaranteed rankings, no mention of redirects, vague answers about who edits content afterward, code hosted only on the agency platform, and no plan for the weeks after launch. Any one of these is worth a direct question. Two or more usually means the agency builds new sites well but has not done many migrations.
When is hiring an agency not worth it?
When the site is small enough to rebuild yourself and has little search traffic to protect, or when WordPress is genuinely doing heavy work such as complex WooCommerce logic, memberships or multi-editor publishing. In the first case a builder and a weekend may be enough. In the second, a migration replaces a working application, and better maintenance is usually cheaper than any agency.
