AI alone is not enough for SEO strategy. AI-generated plans consistently waste resources on low-impact tasks like schema markup, vanity keywords, and micro-optimizations while missing the strategic priorities that actually drive traffic and leads. The best results come from pairing AI’s speed with a human expert’s judgment, an approach sometimes called the “centaur” model.
At TJ Digital, where I’ve spent 16 years in SEO and now use AI in nearly every part of our workflow, I’ve reviewed dozens of AI-generated strategies. They all have the same problems. The plans look thorough on the surface but consistently prioritize the wrong things. Pairing AI with an experienced strategist who knows which recommendations matter (and which ones are noise) is what actually produces ROI.
Can AI Create an SEO Strategy on Its Own?
Technically, yes. Practically, not well.
I had someone reach out to me recently looking for help with SEO. He told me he didn’t need help with strategy because he’d already had Claude create the plan. He just needed someone to implement it.
I told him right away it probably wasn’t a good fit. The primary value we bring to clients is the plan itself. But I was also concerned about what would happen if he followed through on a plan built entirely by AI.
So I asked if I could see it.
This person clearly knew what he was doing when it came to using AI. He was using Claude Opus, which I think is the best model for this kind of work. He’d provided a bunch of context about his brand. But as I looked through the plan, the problems were obvious.
It recommended spending time adding structured data to almost every page of the website, structured data that would have had zero impact on rankings. It recommended targeting keywords that were either irrelevant or far too competitive. And it recommended a bunch of micro performance optimizations that would have taken a ton of resources with little to no impact.
These are the same problems I see in almost every AI-generated SEO plan.
Why Does AI Default to Low-Impact SEO Tasks?
AI models are trained on massive amounts of SEO content, and most of that content is generic checklists. So when you ask an AI to build an SEO plan, it does what it knows: it echoes the checklist.
That means you’ll get recommendations like “add structured data to every page,” “optimize all image alt text,” and “target hundreds of keywords.” These aren’t wrong in a vacuum. They’re real SEO tactics. But for most small businesses, they’re low-priority at best and a complete waste of time at worst.
Take structured data as an example. AI tools love to recommend schema markup as if it’s a ranking factor. It isn’t. Recent testing has shown that large language models often ignore JSON-LD content entirely because it gets broken apart during tokenization. AI search systems learn from human-visible text, not from structured data pasted into your page code. If your content and authority aren’t there, no amount of schema is going to fix that.
The same pattern plays out with keyword research. AI can brainstorm keyword ideas all day, but it can’t judge competitiveness with any accuracy. It might suggest terms that sound relevant but are dominated by massive brands with years of authority. A seasoned SEO would know instantly that a new site needs to focus on low-competition, buyer-intent phrases, not dream keywords.
| AI-Generated Plan (Common) | Expert-Guided Plan |
| Add schema markup to every page | Add schema only where it provides clear context (e.g., FAQ, local business) |
| Target high-volume head terms | Target winnable, buyer-intent keywords with local modifiers |
| Optimize page speed across the board | Fix speed only if pages take 5+ seconds to load |
| Create dozens of blog posts on broad topics | Build high-quality service pages first, then target middle-of-funnel blog content |
| Audit and fix every technical issue | Fix crawlability and indexation issues, skip cosmetic tweaks |
What Happened When a Human Reviewed the AI Plan
Here’s where the story gets interesting. The person I mentioned was smart. After I made a 20-minute Loom video breaking down all the problems with his AI-generated plan, he took the transcript of that video, gave it back to Claude, and asked it to rework the strategy.
The updated plan was actually pretty good. Not perfect, but a massive improvement.
This is the part most people miss. They think that because AI can’t create a strategy on its own, you shouldn’t use AI when developing your strategy. That’s exactly wrong.
How Does the “Centaur” Approach to SEO Work?
Think about chess. There was a period where the best chess players in the world could beat computers. Then computers got good enough to beat the best humans. But for a long time after that, a good chess player working with a computer could beat a computer playing on its own.
That’s where we are with SEO strategy right now. An expert plus AI beats AI on its own, every time.
I don’t play a lot of chess, but my understanding is that computers have now gotten so good at chess that the human just gets in the way. I believe that will eventually be the case for all knowledge work. But we’re not there yet. Not even close when it comes to SEO strategy.
This “centaur” model works because the human and the AI contribute different things. AI brings speed, breadth, and the ability to process large amounts of information quickly. The human brings judgment, context, and the ability to prioritize.
When you’re developing a strategy that determines where you allocate all of your resources, you really don’t want to get that wrong.
What Does AI Actually Get Wrong About SEO?
The problems fall into a few predictable categories.
Chasing volume instead of outcomes. AI tools optimize for keywords without understanding why someone is searching. According to Search Engine Land, over 86% of marketers still heavily edit AI-generated content to add human perspective and accuracy. The “savings” from using AI often get consumed by the time spent fixing its output.
Over-indexing on micro-optimizations. A clean technical foundation is enough for most small sites. You don’t need advanced schema strategies or constant micro-optimizations at the start. You need crawlable pages, fast load times, and content that actually answers what people are searching for.
Missing the strategic context. AI doesn’t know your market position, your budget constraints, or which competitors you can realistically outrank. It can’t tell you that a keyword is technically strong but commercially useless for your specific situation. It won’t push back when something “technically works” but doesn’t work for your audience.
How Should You Use AI for SEO Strategy?
The key is to feed AI rich context so it can produce better output, and then apply human judgment to filter the results.
Start by giving the AI everything it needs: your brand details, target audience, competitors, budget constraints, and any existing data you have. The more specific the input, the more useful the output. A vague prompt gets you a generic checklist. A detailed prompt with real business context gets you something you can actually work with.
This is why transcripts and real conversations are so valuable as AI inputs. The Loom video I recorded for that prospective client wasn’t a polished strategy document. It was me reacting in real time to what I saw. But when that transcript was fed back to the AI, it completely changed the quality of the output. The AI suddenly had access to expert judgment it didn’t have before.
The same principle applies to discovery calls, sales conversations, and even voice memos. The best AI inputs come from people talking naturally about what they know. This is why AI is reshaping the SEO industry so quickly: the practitioners who know how to feed AI the right context are getting dramatically better results than those using AI as a replacement for expertise.
What Should a Good SEO Plan Actually Focus On?
If you’re a small or medium business, the highest-ROI activities are almost always the same:
- Build high-quality service pages targeting bottom-of-funnel, buyer-intent keywords
- Optimize your Google Business Profile
- Earn a handful of relevant, authoritative backlinks
- Create middle-of-funnel blog content targeting questions your customers ask before they buy
- Get your technical basics right (crawlability, mobile experience, load time)
That’s the 20% of effort that drives 80% of qualified traffic. Most businesses fail at SEO not from doing too little, but from doing too much of the wrong things.
An AI plan will often bury these priorities under a mountain of lower-impact tasks. A human strategist puts them front and center.
How Do You Know If Your SEO Strategy Needs Expert Review?
If your plan was generated entirely by AI and you haven’t had an experienced SEO evaluate it, there’s a good chance you’re about to spend time and money on tasks that won’t produce results. Here are some red flags:
- The plan recommends adding structured data to every page without explaining the expected impact
- It suggests targeting keywords with monthly search volumes over 10,000 when you’re a new or small site
- It includes a long list of technical optimizations but no clear priority order
- It doesn’t mention your specific competitive landscape or market position
- It treats all blog content equally rather than prioritizing by buyer intent
FAQ
Does structured data help you rank in AI search?
Not directly. Structured data helps machines parse your content, but it isn’t a ranking factor in Google or AI search. Testing has shown that large language models primarily extract information from human-visible text, not from JSON-LD markup. Schema is worth adding in specific cases (like local business info or FAQ sections), but it won’t compensate for weak content or low authority.
How accurate is AI keyword research?
Not very. AI-generated keyword suggestions are a reasonable starting point for brainstorming, but the search volumes and difficulty estimates are often significantly off. Real keyword tools like Ahrefs or Semrush pull from actual search data. AI models estimate based on their training data. Always validate AI keyword suggestions with real data before building your strategy around them.
Should I stop using AI for SEO entirely?
No. AI is incredibly useful for SEO when you pair it with human judgment. Use it to brainstorm, draft content, analyze data, and accelerate repetitive tasks. Just don’t hand it the steering wheel on strategy without someone experienced reviewing the output.
Ready to get an expert review of your SEO strategy? Contact TJ Digital for a free marketing audit. We’ll tell you exactly what’s worth your time and what isn’t.