Hey Reader,
I'm going to be direct, probably more direct than most newsletters in your inbox.
Because I'm genuinely concerned about something I keep running into, and I'd rather say it plainly than dress it up.
Brands are handing their SEO decisions to AI.
Not using AI as a tool. Not using it to support a strategy. No keyword research. No consideration of site history. No awareness of existing ranking momentum.
Fully handing it the wheel.
No human review of what those changes actually do to the technical foundation underneath. And I'm watching the damage happen in real time.
This isn't a post about being anti-AI. You've read enough of my newsletters to know I use it constantly. It's one of the most powerful tools in my workflow.
This is a post about what happens when a powerful tool gets used without the expertise to wield it correctly. And right now, that's happening at a scale that's going to hurt a lot of brands before they realize what went wrong.
ποΈ Imagine This Scenario
You've spent months building a racecar. The engine is tuned. The suspension is dialed. Every component was chosen deliberately, for this track, this race, this strategy.
Now imagine someone walks into your garage between races and starts swapping parts.
New tires, different compound than what your team chose. Adjusted fuel mix, based on a generic recommendation, not your specific engine build. Moved a few components around, looked fine to them, but they didn't know why those parts were positioned that way in the first place.
The car still looks like a racecar. It still starts. It pulls out of the garage fine, but will it go the distance?
"You won't know something is wrong until you're three laps in at race speed. And by then, the damage is already done."
That's exactly what happens when AI makes uncoordinated, unsupervised changes to a website with an existing SEO strategy, an established ranking trajectory, and a technically built architecture.
The site still loads. The pages are still there. Everything looks fine, until it doesn't.
What AI Genuinely Cannot Know About Your Site
This is the part I need you to really sit with, because this is where the real danger lives.
AI tools, no matter how sophisticated, operate without context, which is absolutely critical for making sound SEO decisions. Here's what they don't have access to:
β Your site's ranking history and momentum
Your site has pages that have been climbing steadily for months. There are keyword positions you've been building toward. There's a trajectory in Google Search Console that represents real, compounding work. AI doesn't see that history. It sees the page in front of it, not the journey that got it there. A content change that looks neutral in isolation can interrupt a ranking signal that took six months to build.
β Your intentional site architecture
The structure of your site, how pages link to each other, how categories are organized, and how authority flows through internal links was built with a plan. That plan exists in someone's head, in a strategy doc, or in years of accumulated decisions. AI has none of that context. It will reorganize, rewrite, and restructure based on what it finds logical in the moment. Not what's logical for your specific SEO strategy.
β Your keyword strategy and cannibalization risks
Keyword research is not a one-time event. It's an ongoing map of which pages own which terms, which topics are intentionally separated to avoid competing with yourself, and which gaps you're actively working to fill. AI doesn't know your keyword map. It will happily create content that cannibalizes your own top-performing pages, and you won't see the damage until rankings start dropping on pages you weren't even paying attention to.
β Your crawl budget and indexing priorities
Not every page on your site is equal in Google's eyes, and your crawl budget is a finite resource. There's a reason certain pages get crawled frequently and others don't. AI-driven content additions, page restructuring, and bulk edits can throw crawl patterns into chaos, dilute indexing priority on your most important pages, and create technical debt that takes months to clean up.
β Your niche-specific ranking signals
Every industry has nuances in how Google evaluates expertise and authority. A healthcare site operates under completely different E-E-A-T requirements than an e-commerce site. A local service business has entirely different technical priorities than a national content publisher. AI generates advice for the average case. Your site is not the average case.
π The hard truth:
AI sees a snapshot. Your SEO strategy is a film.
Every change made without understanding the full story is a bet placed without knowing the odds.
And the house always wins eventually.
What I Actually See in Search Console After This
When I audit a site that's been running on uncoordinated AI-driven changes, the patterns are consistent. Every time.
|
- Impressions slowly dip, clicks in freefall: pages still appearing in results but losing click-through because titles and meta descriptions were rewritten without testing.
- Ranking volatility on previously stable pages: positions swinging 10β30 spots week over week because content changes disrupted established relevance signals.
- Keyword cannibalization across multiple pages: AI-created content now competes directly with existing top performers.
- Crawl coverage errors are increasing: new pages are added without proper internal linking, and old pages are modified in ways that confuse crawlers.
- Site speed regressions: code or content additions that weren't evaluated against Core Web Vitals performance.
None of this shows up immediately. That's what makes it so dangerous. You make the changes, the site looks fine, and three weeks later you're staring at a GSC report wondering what happened.
By the time the damage is visible, it's already compounded.
π The Checklist: What to Verify Before Any AI-Driven Change
I'm not telling you to stop using AI. Use it. But before any AI recommendation touches your live site, run it through this:
π BEFORE YOU PUBLISH, RUN THIS CHECK:
β KEYWORD IMPACT
- Does this change affect a page that's currently ranking for target keywords?
- Could this new content cannibalize an existing page targeting the same or similar terms?
- Has proper keyword research been done for this specific topic, not just AI-suggested terms?
β ARCHITECTURE & STRUCTURE
- Does this change fit within the intentional site architecture, or does it create an orphaned page?
- Are internal links updated to reflect this addition or edit?
- Does this disrupt any existing content clusters or topical authority groupings?
β TECHNICAL HEALTH
- Has this page been checked for Core Web Vitals impact after changes?
- Are title tags and meta descriptions still optimized, or did AI rewrite them without review?
- Have structured data/schema elements been preserved or updated correctly?
- Is the page properly indexed, or does it need a crawl request after significant changes?
β HISTORY & MOMENTUM
- Does this page have existing ranking history that could be disrupted?
- Has performance in GSC been checked before and flagged as a baseline?
- Is a human SEO strategist aware that this change is being made, and have they signed off?
π The rule:
If AI recommended it and a human with SEO expertise hasn't reviewed it, it doesn't go live.
Full stop.
|
Use AI. Don't Surrender to It.
I'll say it one more time because it matters: AI is a powerful tool. In the right hands, with the right oversight, it accelerates everything.
But a racecar without a driver isn't fast. It's dangerous.
Your website has a strategy. It has a history. It has momentum. It has technical decisions baked into its architecture, each with a specific reason.
AI doesn't know any of that. It can't. And every unsupervised change it makes is a roll of the dice with rankings you've worked hard to build.
The brands that win in the long term are the ones that use AI for speed and human expertise for judgment. Not one or the other. Both, in the right order.
β