The experiment

Starting from zero — no existing audience, no domain authority, no social following — I built and launched BuzzRiding using AI as the primary content engine. Claude wrote the briefs, the articles, the SEO metadata, the social posts, and the newsletter copy. My role was to direct, review, and occasionally rewrite the intro.

This is the honest account of what that produced.

📋 Experiment Parameters

Duration: 30 days. Articles published: 8. AI tool: Claude (claude.ai). Human review time per article: 15–20 minutes. Total active time on content: approximately 3 hours over the month.

The results

8
Articles published
3hrs
Total active time
€0
Total cost
~0
Organic traffic (month 1)

The traffic number needs context. Zero organic traffic in month one is expected — not a failure. A new domain with zero authority takes 3–6 months to appear in Google results, regardless of content quality. The articles are indexed. The clock is running.

What the AI did well

Speed was extraordinary. From keyword research brief to published-ready article in under 45 minutes. For eight articles, that's approximately 6 hours of AI work compressed into 3 hours of my active time. No human writer works at that pace.

Consistency was better than expected. Keeping a consistent brand voice across 8 articles written in separate sessions is genuinely hard for a human. Claude maintained the BuzzRiding voice — friendly, data-led, jargon-free — across every piece with minimal correction.

SEO structure was solid out of the box. H2 structure, FAQ sections, meta descriptions — all done correctly on the first pass. No SEO specialist needed.

What the AI did badly

Intros were the weakest element. AI-generated article openings are recognisably formulaic. They state the problem, promise to answer it, and get to the point. That's correct structure — but it's also predictable. Every intro needed a human rewrite.

Specificity required prompting. Left to generate freely, Claude produces accurate but generic content. The quality gap between a generic prompt and a detailed brief with specific data points, examples, and angle guidance is enormous. The workflow works — but only if you invest in the brief.

No genuine experience. The articles are well-structured and informative. They don't have the lived-experience texture that the best content writing has. That's a real limitation — one that becomes more important as Google's E-E-A-T signals mature.

Would I do it again?

Yes. Without hesitation. The alternative — writing 8 articles manually in 30 days while building everything else from scratch — was not realistic. AI made the project possible.

The correct framing is not "AI vs. human writing". It's "AI-assisted writing vs. no content at all". At the zero-budget, solo-operator stage, the comparison is obvious.

The quality ceiling is real. As the brand matures and develops genuine audience data and real experiments to reference, the content will need to develop a thicker layer of authentic human experience. That's a month 3+ problem.

Frequently Asked Questions

Can you really build a blog entirely with AI?
Yes — with important caveats. AI handles structure, research synthesis, and first-draft writing extremely well. What it doesn't provide is genuine first-hand experience, original data, or the specific insights that come from actually doing the work being written about. The best AI-assisted blogs combine AI efficiency with human insight at the review stage.
Will AI-written content rank on Google?
Google's position is that it evaluates content quality, not the method of production. AI content that is accurate, well-structured, and genuinely useful can rank. AI content that is thin, generic, or produced without a clear audience in mind typically does not. The workflow matters more than the tool.
How long does it take to write an article with AI?
Using the BuzzRiding workflow: keyword research and brief (10 min AI + 5 min human direction), article draft (10 min AI generation), human review and edit (15 min), SEO metadata (5 min AI). Total: approximately 45 minutes per article from brief to published-ready.