“Why can’t we just use AI to do it?”
Whether you’re on the brand or agency side of SEO, I’m guessing you’ve heard some version of this from an exec or a client with little knowledge of AI tools, SEO principles, or both.
I’ve been asked that question multiple times because the other party saw or heard about modest success from LLM-generated content that got some clicks and impressions.
My answer: because thousands of LLM-produced pieces of content do not a successful SEO program make.
This article dives into the human and AI roles in today’s SEO landscape, including:
What people are getting wrong about AI and content.
What AI can and can’t do for SEO campaigns.
What an expert can tackle with AI tools.
The North Star of 2025 SEO (as I see it) and why you need humans to reach it.
(Note: No LLMs were used to write this article.)
What people are getting wrong about AI and content
When people ask, “Can we just have AI write 1,000 blog posts?,” they assume there’s a linear progression.
For instance, if a blog post gets 100 visits/month, won’t 1,000 blog posts get 100,000 visits?
No, that’s not the way SEO works. It’s not a linear discipline.
More importantly, that approach means you’re just putting crap out there. You’re essentially using AI to build your own content farm of stale, repetitive language.
There’s no value for the user or positive affinity for the brand.
Now, you could use AI tools and strategic prompts to quickly create a solid base for a piece of content, then apply human editing and a unique POV.
In most cases, that’s faster than the content process was before AI, and it’ll produce much better content than 1,000 LLM-produced pieces, but it still requires human input.
In short, forget about spamming Google with a ton of poor LLM content. Your users won’t read it, and ultimately, it won’t do anything beyond maybe inflating your vanity metrics.
And, crucially, Google won’t like it.
Whenever Google deals with an explosion of people doing the same (easy) thing to game the system, you want to zig while others are zagging.
Don’t be part of the problem that triggers – and gets wiped out by – a huge algo update.
Dig deeper: 3 ways to use AI for SEO wins in 2025
What AI can and can’t do for SEO campaigns
Along with being unable to produce differentiated content, AI is being asked to do things like “come up with keywords” or “do internal links” on its own.
If you’re just having AI look at your site and update links without careful QA, you’ll just end up with a lot of crappy internal links.
It’s the same thing with keywords: you might get a huge list, but lots of them will have low volume, be barely relevant, or be straight-up garbage.
Anytime someone says, “Let’s just use AI for [task],” try it once, gauge the output and the time needed to bring it up to anything resembling human baseline, and you’ll have a more nuanced answer.
On the other hand, there are a few proven use cases for AI in SEO – and while they still involve human input, they’re big time-savers that free up the experts to address more strategic initiatives.
For instance, if you have good source data and/or good, well-substantiated original thoughts, AI is great for remixing them into something organized and usable.
Let’s say you conduct a thorough interview with a solutions engineer. AI can highlight, categorize, and synthesize the most salient parts of the interview, leaving you to QA the output and layer in your own voice.
Not only does this save you time, it helps surface patterns in big data sets that you might never have spotted on your own – or at least nowhere near as quickly.
Dig deeper: How to optimize your 2025 content strategy for AI-powered SERPs and LLMs
Get the newsletter search marketers rely on.
If you approach AI tools with the right expectations, they can be incredibly powerful.