Blogging Titan research series
AI tools and AI claims, tested on tasks where we already knew the right answer.
The standard for this series. Prompts, settings and dates are published with every test. We test through normal product use, grade against a rubric set before the run, and report the cases where the tool did well as plainly as the failures.
In this series
- Which UK Businesses Does AI Recommend? What 160 Google AI Mode Answers Show
- Can AI Give UK Businesses Accurate Marketing Prices? We Re-Ran 220 Answers
- We Gave 5 AI Tools (8 Models) the Same Blog Brief. Here Is Every Word They Wrote, Scored
- AI Traffic Converts 1.3x Better Than Organic. Or 4.4x. Or 23x. We Traced Every Claim to Its Source.
- Schema Markup Won’t Win You AI Citations: We Checked 138 Pages in Google’s AI Overviews
- The First-Person Fallacy: Does ‘I Tested’ Language Win AI Citations?
- What AI-Citable Content Looks Like: We Ran the Answer-First Test on 152 Blog Posts (2026)
- Best AI Content Detectors 2026: 8 Tested on Text We Knew the Answer To
The other five series
The Internet Experiment · The Website Autopsy · Hidden Economics of the Internet · The Internet Investigation · The Next Internet
Every series sits inside Original Research, alongside our named findings, how we test and our editorial policy.