MCP vs. AI Link Suggestions: What’s the Difference?

Short answer: MCP doesn't replace AI Link Suggestions, it sits on top of it. They're not two competing engines, they're the same intelligence reached two different ways.

The engine vs. the front door

AI Link Suggestions is the engine. It runs continuously in the background: reading every post, generating keywords, classifying intent and funnel stage, computing OpenAI embeddings, and scoring which internal links make sense. This is where the actual work happens.

MCP is a front door, not a second engine. When you ask Claude or ChatGPT to suggest internal links for a post, the suggest_internal_links tool doesn't call OpenAI itself, it reads whatever the AI Link Suggestions engine already computed and hands it back in plain English. If the engine hasn't processed a post yet, MCP has nothing to give you either.

Side by side

So is one obsolete?

No. Turn off the AI Link Suggestions engine and the MCP suggestion tool goes empty; there's nothing left for it to read. They depend on each other:

• Keep using the dashboard when you want to systematically review and apply suggestions across your site.
• Use MCP when you want to ask a quick question from outside WordPress, like "which pages should link to my new post?", without opening wp-admin at all.

The one thing to know

Because MCP is just a window into existing data, connecting an AI assistant doesn't cost you anything extra in OpenAI usage for link suggestions specifically. The cost you already pay for AI Link Suggestions processing is the same either way; MCP just gives you another way to reach the results.