Bankrate.com uses AI-generated content extensively. That content ranks highly on Google and generates about 125,000 organic visitors per month on AI-assisted pages.

Meanwhile, the December 2025 core update specifically targeted AI content quality and sites relying on mass-produced, unedited AI text saw traffic decimated.

Same technology. Opposite results. The difference is how it’s used.

AI for SEO content production is neither magic bullet nor guaranteed penalty. It’s a tool that can scale quality content or scale garbage. The outcome depends on the implementation.

Google’s Actual Position

Let’s clear up the confusion about what Google allows.

Google’s official position is clear: “Our systems don’t care if content is created by AI or humans. We care if it’s helpful, accurate, and created to serve users rather than just manipulate search rankings.”

That’s not a loophole. It’s a higher standard.

Google doesn’t penalize AI content. They penalize unhelpful content. The problem is that most AI content, produced carelessly, happens to be unhelpful. Not because AI wrote it, but because nobody ensured it was actually good before publishing.

As one analysis puts it: AI content is not inherently bad for SEO. Lazy AI content is fatal to SEO. In 2026, Google’s ranking systems are sophisticated enough to differentiate between a helpful resource assisted by AI and derivative content designed solely to capture keywords.

The watershed moment came with the massive core updates of 2024, which targeted “scaled content abuse” - the practice of churning out thousands of low-quality, AI-generated pages. Sites that relied on mass-produced, unedited AI text saw their traffic decimated.

The rule is now clear: automation without curation is a liability.

The Search Landscape Has Changed

Understanding why AI SEO content is tricky requires understanding how search itself has changed.

AI Overviews Dominate

Google’s AI Overviews now appear in more than 50% of all search results. Just ten months ago, that figure was 25%. The feature has reached 2 billion users.

The growth is dramatic. Between March 13-27, 2025, AI Overviews grew 528% for entertainment queries, 387% for restaurant queries, and 381% for travel queries. Queries with eight words or more are 7x more likely to get an AI Overview.

Click-Through Rates Are Declining

CTRs for Google’s top-ranking search result have declined from 28% to 19% as AI Overviews take the most valuable real estate. Position two experienced an even steeper decline, with CTRs falling 39% from 20.83% to 12.60% year-over-year.

Across positions 1 through 5, there was an average CTR decline of 17.92%.

Gartner predicts a 25% drop in search traffic by 2026 due to AI chatbots and similar systems answering queries directly.

New Traffic Sources Are Emerging

While traditional search traffic declines, AI referral traffic is growing. Traffic from LLMs rose 527% comparing January-May 2024 with the same period in 2025, from about 17,000 to 107,000 sessions. Some sites now report over 1% of total sessions from ChatGPT, Perplexity, and Copilot.

AI traffic is still relatively small - only 0.1% of total web traffic - but early research shows AI referral visitors convert better than traditional search visitors. And it’s growing 165x faster than organic search traffic.

The implications: success in 2026 means optimizing for AI citations alongside traditional rankings.

What’s Actually Working

Let’s look at documented wins.

Xponent21: 4,162% Organic Growth

Xponent21 achieved 4,162% organic traffic growth in roughly a year, accumulating over 10.5 million impressions and 20,100 total clicks. On their best day, they hit 168,337 impressions, with more than 5% of traffic coming from AI search agents like Perplexity and Bing Chat.

Their approach: highly targeted content with genuine expertise, optimized for both traditional rankings and AI citations.

Bankrate: AI at Scale With Quality

Bankrate.com uses AI extensively and generates about 125,000 organic visitors monthly from AI-assisted pages. They rank highly on Google for both primary and long-tail keywords.

Their approach: AI for efficiency, human experts for accuracy and depth. Financial content requires expertise that AI cannot fake. They provide it.

STACK Media: 61% Traffic Increase

STACK Media achieved a 61% increase in website visits and a 73% reduction in bounce rates using AI-driven tools with BrightEdge.

Their approach: AI for analysis and optimization, human judgment for content strategy.

The Common Thread

Every success story shares characteristics:

AI as assistant, not author. AI handles research, outlines, drafts, optimization. Humans provide expertise, editing, quality control.

Genuine E-E-A-T signals. Experience, Expertise, Authoritativeness, Trustworthiness aren’t optional. AI cannot possess “Experience” - it cannot test a product, visit a destination, or interview a client. Content that ranks demonstrates human involvement.

Depth over volume. The strategies that worked in 2018 - publishing moderate-quality content targeting keyword volume - no longer suffice in 2025. Success requires content depth that AI cannot replicate.

What’s Getting Penalized

The December 2025 core update made it clear what doesn’t work.

Mass-Produced AI Content Without Editing

Sites publishing high volumes of content without substantial depth, unique insights, or genuine expertise are seeing dramatic visibility drops. This includes content farms, sites using excessive automation, and platforms that spread content across large networks without adequate editorial oversight.

The pattern: generate, publish, hope. No expertise. No editing. No value.

Thin Content on Important Topics

Affiliate sites without first-hand testing, thin e-commerce pages, and content farms saw the biggest declines. Sites with niche authority, strong engagement, and experience-led content were more stable or gained visibility.

YMYL topics (Your Money, Your Life - health, finance, legal) face the highest scrutiny. AI content on these topics without expert verification is especially risky.

Keyword-Stuffing at Scale

Google’s dealing with an explosion of AI-generated content that’s technically correct but lacks the depth that comes from actually doing the work. Rather than try to detect AI (which is getting harder), they’re rewarding markers of genuine human expertise.

The signal Google looks for: did someone who knows this topic actually create this, or did someone just optimize for keywords?

The E-E-A-T Challenge

Experience, Expertise, Authoritativeness, Trustworthiness - Google’s primary filter for quality.

AI has a fundamental problem here. It can compile information. It cannot have experience. It can summarize expertise. It cannot possess it.

This creates a ceiling on pure AI content. To rank in 2026, content must demonstrate human involvement.

What that looks like practically:

First-party data and research. Original surveys, studies, data analysis. AI can’t generate these.

Personal experience markers. “I tested this” or “In my experience” with specific details AI couldn’t know.

Expert credentials displayed. Author bylines with verifiable expertise. Real names, real backgrounds.

Original media. Photos, screenshots, diagrams you created. Not stock images or AI-generated filler.

Specific, unique insights. Observations that come from actually working in the field.

The Hybrid Workflow That Works

Based on what’s succeeding and failing, here’s a production workflow.

Stage 1: Strategic Planning (Human)

Before touching AI:

• Identify topics where you have genuine expertise
• Research what’s ranking and why
• Find angles competitors haven’t covered
• Define what unique value you can add

AI cannot do strategy. It can suggest keywords. It cannot know whether you can actually create authoritative content on that topic.

Stage 2: Research and Outlining (AI-Assisted)

Use AI for:

• Gathering background information
• Identifying subtopics to cover
• Creating initial structure
• Finding questions people ask

Don’t use AI to: tell you what to say. Use it to organize what you already know.

Stage 3: First Draft (AI-Assisted)

AI content wins when used as an efficiency engine within a human-led strategy. Using AI to structure outlines, analyze data, or draft sections that are then heavily edited by experts allows for scale without sacrificing quality.

The draft is a starting point. It will not be publishable as-is.

Stage 4: Expert Enhancement (Human)

This is where the real work happens:

• Add your actual expertise and experience
• Include original data, examples, insights
• Verify every fact, stat, and claim
• Add E-E-A-T signals throughout
• Make it sound like you, not like AI

A good rule: if a section could have been written by anyone, rewrite it until it couldn’t.

Stage 5: Quality Control (Human)

• Fact-check everything
• Verify links work and are accurate
• Read aloud for voice consistency
• Check against competitor content
• Ensure genuinely helpful, not just comprehensive

Stage 6: Optimization (AI-Assisted)

AI helps with:

• Schema markup generation
• Meta description drafts
• Internal linking suggestions
• Technical SEO checks

These are mechanical tasks where AI excels.

The Scale-Quality Tradeoff

Here’s the uncomfortable truth: the concepts of content quality and scale are fundamentally opposed. It’s not possible to create high-quality content at scale unless you have a massive team of subject matter experts and content writers and an enormous content budget.

AI doesn’t change this equation. It changes where the bottleneck sits.

Before AI, the bottleneck was writing time. You couldn’t produce content faster than your writers could type.

With AI, the bottleneck becomes expertise and editing. You can generate drafts instantly. But making those drafts publishable still requires human knowledge and judgment.

Realistic Production Rates

If quality AI-assisted content takes 3-4 hours of human work per piece (research, expertise addition, editing, fact-checking), you haven’t eliminated labor. You’ve shifted it.

What AI enables:

• Better first drafts to start from
• More content options to choose from
• Faster research and organization
• Consistent formatting and structure

What AI doesn’t enable:

• Publishing good content without human expertise
• Scaling without proportionally scaling editing capacity
• Ranking thin content through volume

Plan accordingly.

What’s Coming

The search landscape continues to shift.

76.1% of URLs cited in AI Overviews also rank in the top 10 of Google search results. Traditional SEO remains the foundation for AI citation success. 92.36% of successful AI Overview citations come from domains already ranking in the top 10.

This suggests: optimization for AI search is an extension of traditional SEO, not a replacement. Sites ranking well traditionally are the same sites getting cited by AI.

Case studies and pricing pages are now the best content types to drive traffic in the age of AI. Top funnel content (what is, how-tos, guides) saw massive drops in the past two years.

The implication: AI handles informational queries directly. Traffic flows to content that demonstrates specific experience and serves decision-making stages.

The Responsible Path Forward

AI SEO content production isn’t about gaming algorithms. It’s about efficiently creating content that genuinely helps users.

Around 89% of small business owners and marketers use AI for content marketing and SEO. 68% see higher content marketing ROI with AI. This isn’t going away.

The responsible approach:

Use AI for efficiency, not replacement. Faster research, better organization, stronger starting points. Not automated publishing.

Maintain expert oversight. Every piece needs someone who actually knows the topic reviewing it.

Add genuine value. Original insights, data, experience that AI cannot provide.

Stay transparent. Don’t claim AI content is from your experience when it isn’t.

Monitor performance. Track how AI-assisted content performs vs. human-written. Adjust based on data.

The sites succeeding with AI content aren’t trying to trick Google. They’re using AI to help humans create more helpful content, faster.

The sites failing are trying to automate expertise. That doesn’t work because expertise isn’t automatable.

AI won’t write your way to SEO success. But it can help you get there faster, if you bring the expertise and the quality standards yourself.