TL;DR
AI answers are absorbing informational clicks. Treat LLMs as a new distribution layer – design content to be cited, credited, and convertible from inside summaries.
(Wikimedia)
What happened
Wikimedia’s latest update reports human pageviews fell ~8% year‑over‑year after cleaning out bot noise. Their reading is straightforward: generative AI in search now answers many queries directly, often drawing on Wikipedia, while social video siphons casual lookups. The result – fewer sessions reaching source pages even when those pages underpin the answer.
(Wikimedia)
Why this matters for LLM SEO
Classic SEO optimized for ranking positions and blue‑link CTR. LLM SEO optimizes for inclusion in the answer. If the answer layer becomes the default for “how/what/why” queries, then your KPIs must expand beyond rank: citations in AI Overviews, mentions in ChatGPT/Gemini answers, and assisted conversions from summary‑origin traffic. Publishers already fear what Italy’s FIEG calls a “traffic killer,” and regulators may push UX toward stronger attribution – but you shouldn’t wait for policy to save your funnel.
(Mediapost)
The playbook to adopt now
- Reference‑grade pages – Lead with a concise, fact‑dense summary, then expand. Use clear headings, tight paragraphs, and canonical definitions.
- Liftable passages – Craft 2–3 sentence snippets that cleanly answer the core question. Mark them up with FAQ/HowTo/Article schema where appropriate.
- Entity clarity – Disambiguate names, dates, metrics, and relationships. Use consistent terms so retrieval systems resolve you correctly.
- Evidence and assets – Publish small datasets, methods, or checklists. LLMs favor concrete artifacts they can cite.
- Attribution plumbing – Ensure titles, bylines, org names, and licensing are machine‑readable. Maintain a /credits or /about‑data page that explains provenance.
- Measure AI visibility – Track when your brand appears in AI answers for priority intents. Build an “AI mentions” log alongside rank tracking.
- Conversion from summaries – Offer canonical URLs with jump links, TL;DRs, and downloadable artifacts. If a user does click, make it count.
What to watch
- Regulatory pressure in the EU – Investigations into AI Overviews could reshape summary UX – more obvious links and source prominence would reward citation‑friendly sites. (MediaPost)
• Ranking turbulence as AI surfaces evolve – Monitor volatility alongside AI Overview presence on SERPs to separate algorithmic shifts from UI‑driven CTR changes. (Search Engine Roundtable)
Bottom line: Treat AI answers as a distribution channel. If you become the source model’s trust, you’ll keep attention – and revenue – even as clicks consolidate.
