Running content operations across 30 client sites gives you a different perspective on algorithm updates than you get from managing a single site. You see patterns across different niches, different domain authorities, different publishing frequencies, and different levels of editorial investment. When a major update hits and some accounts take hits while others do not, you can start to see what the differentiating variables actually are rather than just telling a story about your own one case.
We use AI tools extensively in our content production. We have for about two years. The question we keep getting from new clients is whether Google is banning AI content or moving toward it. Here is what managing content across dozens of sites has taught us about what Google actually cares about.
The Pattern Across Client Sites After Major Updates
After the February 2026 core update, we had six client sites that took meaningful traffic drops of 15 percent or more. We had twenty-four that either held steady or gained ground. All thirty sites were using AI tools in some form for content production. If AI use were the variable driving the drops, you would expect a much more even distribution of negative outcomes. That is not what we saw.
Looking at the six sites that dropped, the common threads were not about AI. They were about content strategy: two of them were in highly competitive niches where their domain authority was significantly outmatched by the sites they were competing against; two had been publishing at very high volume relative to their editorial capacity, meaning the review process was thin or inconsistent; and two had large content libraries from before they became clients with us that contained a significant amount of old, thin content we had not yet had time to clean up. The AI production method was present in all thirty. The specific strategic problems were present only in the six that struggled.
What Our Editorial Process Looks Like and Why It Matters
We produce first drafts with AI tools for about 70 percent of the content we create across client accounts. Every draft goes through a specific editing process before it is published: a fact-check of any specific claims that can be verified; a tone adjustment pass to match the client’s established voice; the addition of at least one specific detail per article that reflects genuine expertise in the field; often this means having someone with subject matter knowledge review the section covering the most technical aspect of the topic); and a final check against the intent of the target keyword to confirm the article is actually answering the specific question someone searching that term is likely to have.
That process takes 20 to 35 minutes per article depending on the topic complexity. It is not fast. But it is the difference between the content that comes out of our workflow and the content that shows up when someone uses an AI tool with minimal oversight. The editing step is where the E-E-A-T signals get added, and those signals are what Google is measuring when it evaluates whether content deserves to rank for a given query.
The Clients Who Panicked and Made the Wrong Move
After the February update, we had two client contacts who read alarming posts about Google cracking down on AI content and decided to pause all AI-assisted content production entirely for two months. They wanted to switch to fully human-written content while the situation “stabilized.” We tried to talk them out of it. One listened; one did not.
The one who paused their AI-assisted workflow while their editorial process remained exactly the same saw no meaningful change in their ranking trajectory. The sites around them continued growing. Their publishing cadence slowed. Their topical authority development slowed proportionally. Three months later their traffic was lower than it had been because they had published significantly less content, not because of any change in how Google evaluated what they were producing.
The lesson from that experience is one I share with every client who raises the AI content question: the variable Google is measuring is not how your content was produced. Changing your production tool without changing your content quality standards produces no change in your ranking outcomes.
What Google’s Quality Raters Actually Look For
Google’s Quality Rater Guidelines are publicly available and worth reading directly rather than through summaries. What quality raters are instructed to assess is not whether content was produced by AI. It is whether the content demonstrates expertise and trustworthiness, whether it fulfills the intent of the person who would search the query, whether it contains original information not readily available elsewhere, and whether it reflects the kind of effort that suggests someone who genuinely knows the subject prepared it.
Unedited AI content fails most of those criteria not because of anything intrinsic to AI but because unedited output lacks the specificity, the nuance, and the experiential grounding that demonstrates genuine engagement with a topic. Those things can be added by a skilled editor. When they are added consistently, the content meets the quality rater criteria regardless of how the draft was initially generated.
When they are not added, the content does not meet those criteria regardless of how many human hours were spent producing it. We have seen plenty of fully human-written content fail quality standards. The production method is the least important variable in the equation.
The Conversation We Now Have With Every New Client
When a prospective client comes to us worried about AI content and Google, we redirect the conversation as quickly as possible to the actual question: is your content strategy producing pages that are genuinely better than what your competitors are publishing for the same terms? Not different. Not more of it. Genuinely better in the specific ways that matter to someone who searched that query with a real need.
If the answer is yes, the AI question is irrelevant. If the answer is no, the problem existed before AI entered the picture and will persist after any change in production method. Getting that question answered accurately requires the kind of competitive content analysis and quality evaluation that tools like SEOZilla provide. That analysis is where we spend our time with clients, not on discussions about which content tool is safe to use. Those discussions are a distraction from the actual work of building a content strategy that earns organic growth.
Hamza 
Blitz