Content formats that get cited in Gemini and AI Mode
Which page types earn links in AI Mode, AI Overviews and Gemini: non-commodity content, comparisons, official task pages and first-hand evidence.
Google's 2026 guidance on generative AI search reduces content strategy to one idea: create non-commodity content. Commodity content, like "7 Tips for First-Time Homebuyers", repeats common knowledge anyone could write, or a model could generate. Non-commodity content, like "Why We Waived the Inspection & Saved Money: A Look Inside the Sewer Line", brings first-hand experience or expert takes. AI Mode and AI Overviews already know the common knowledge. They link to pages that add something. This lesson turns that guidance into page formats B2B and ecommerce marketers can ship, and lists what Google says not to bother with.
- Google: unique, non-commodity content "will likely influence your website's presence in generative AI search in the long run more than any of the other" practices.
- Formats that add something: first-hand guides, original data, honest comparisons, detailed integration and implementation pages, and clear official task pages.
- Organise with paragraphs, sections and clear headings, and support text with high-quality images and video.
- Google says there's no need to chunk, no ideal page length, and no need to rewrite for AI.
- Mass-producing pages for fan-out variants breaches the scaled content abuse policy.
Checked 5 November 2026 (CET) against Gemini Apps Help, the Gemini release notes, Google Workspace Help, Google Search Central, Search Console Help, the Gemini API docs and the Google blog posts cited in Sources. Google renames and moves Gemini features often, so confirm labels and limits in your own account. "Our view" marks MentionX opinion. Ledgerline is a fictional company used for illustration.
01What does Google mean by non-commodity content?
| Commodity (adds little) | Non-commodity (adds something) |
|---|---|
| "What is AP automation?" restating definitions | "We automated 40,000 invoices: what broke in month one" |
| "10 benefits of AP automation" | "AP automation ROI from 212 mid-market customers (2026 data)" |
| "Best AP software" listicle copied from review sites | Honest comparison with test criteria, screenshots and where you lose |
| Generic integration logo grid | "NetSuite integration: fields synced, setup steps, limits" |
| AI-rewritten competitor content | Expert commentary from your controller-in-residence |
Google's guide says: "Don't just recycle what others on the internet have already said, or could easily be produced by a generative AI model."
02Which formats tend to get cited?
Our view, based on MentionX audits and consistent with Google's guidance:
| Format | Why it works with RAG and fan-out | Must include |
|---|---|---|
| Integration and implementation pages | Answer specific fan-out queries ("X with NetSuite", "implementation time") | Steps, fields, timelines, limits |
| Pricing page with plain numbers | Pricing is a common fan-out query; official pages may get a dedicated section | Plans, prices, what's included, billing terms |
| Original research | Unique facts other pages can't supply | Method, sample, date, downloadable data |
| Comparison pages (you vs Competitor A) | Comparison queries are frequent in AI Mode | Fair criteria, where the competitor wins, date checked |
| First-hand guides | Experience is what AI can't generate | Real screenshots, numbers, named author |
| Customer stories with metrics | Specific outcomes | Company type, before and after, quote |
| Glossary pages for niche terms | Clear definitions grounded in your expertise | Definition first, then examples |
| Video walkthroughs | Google says to support text with high-quality images and video | Transcript or summary on the page |
03How should you structure a page?
Google recommends writing for people, with paragraphs, sections and "headings that provide a clear structure". A template we use for product-adjacent pages:
- H1 that states the topic plainly.
- Answer first: two or three sentences that resolve the main question.
- Key facts table: price, integrations, limits, dates.
- Question-style H2s that match sub-questions buyers ask.
- Evidence: screenshots, data, quotes, with dates.
- Who it's not for: honest limitations.
- Author and review date, with a link to the author page.
See Headings, lists and tables LLMs prefer and Answer-first content.
04What should you not do?
From Google's optimisation guide:
| Tactic | Google's position |
|---|---|
| Chunking content into tiny pieces | Not required; "There's no ideal page length" |
| Rewriting content in a special style for AI | Not needed; systems understand synonyms |
| llms.txt or AI text files | Ignored by Google Search |
| Pages for every fan-out query variant | Violates the scaled content abuse policy if done mainly to manipulate |
| Inauthentic mentions | Targeted by spam systems |
| Trusting tools claiming "internal" Google metrics | "No third-party tool has access to our internal ranking or AI systems" |
Our view: the scaled content line matters for SaaS teams with programmatic pages. A page per integration with real setup details is useful. A page per "AP automation for [city]" with swapped words isn't, and it risks your whole site.
05How do images and video help?
Google says many people appreciate images and video, and AI features increasingly show related links with thumbnails (15 October 2026 update). Practical steps:
- Use original screenshots and diagrams rather than stock images.
- Add descriptive alt text and file names.
- Host product walkthrough videos and embed them with a written summary.
- Make sure images are crawlable, not loaded only after user interaction.
06How does the 15 October 2026 update change priorities?
| Change | Content implication |
|---|---|
| Inline links appear higher and more often | Specific, quotable claims win links |
| Test of a dedicated section for official pages | Clear official pricing, login, support and integration pages |
| Related links with thumbnails | A strong lead image on key pages |
| Clickable creator and publisher profiles | Consistent author and brand profiles |
| Preferred Sources highlighted in news carousels | Publishers: promote "add as a preferred source" |
07Worked example: Ledgerline
Ledgerline is a fictional accounts-payable software company. Details are illustrative.
Ledgerline audited 40 blog posts against Google's commodity test.
| Content group | Posts | Commodity? (illustrative) | Action |
|---|---|---|---|
| "What is" definitions | 14 | Yes | Merged into 4 glossary pages with original examples |
| Listicles ("benefits of") | 9 | Yes | Rewritten with customer data or retired |
| Customer stories | 6 | No | Added metrics tables and dates |
| Integration guides | 5 | Partly | Expanded with setup steps and limits |
| Annual AP benchmark | 1 | No | Refreshed with 2026 data and method |
| Comparisons vs Competitors A-D | 4 | Partly | Added test criteria and "where they win" |
Walkthrough
- Each post was scored on one question: could someone write this without our data or experience?
- Commodity posts were merged, rewritten with first-hand material, or redirected.
- The benchmark got a new method section, a downloadable CSV and a named analyst author.
- Comparison pages got a dated test table and an honest "choose Competitor B if" section.
- Over the next quarter, the Search Console Generative AI performance report showed impressions moving to the benchmark and integration pages (illustrative), which the team tracked monthly.
08Common mistakes
- Mass-producing variants for fan-out queries.
- Publishing AI-written commodity posts and expecting citations.
- Hiding prices that buyers and AI both look for.
- Comparisons that never admit a weakness.
- No dates or authors on factual content.
09Try it yourself
Take your five most-visited blog posts. For each, answer: what does this page say that nobody else could? If the answer is "nothing", list one first-hand element (data, screenshot, customer quote, test) you can add this month.
10Quick quiz
1. What's Google's example of non-commodity content?
Answer: "Why We Waived the Inspection & Saved Money: A Look Inside the Sewer Line".
2. Is there an ideal page length for AI features?
Answer: No.
3. Why avoid pages for every fan-out variant?
Answer: Doing it mainly to manipulate results violates the scaled content abuse policy.
11What should you check before the next lesson?
- A commodity audit of your top content.
- At least one original data or first-hand page planned.
- Official task pages that are clear and current.
Previous: Structured data, entities and the Knowledge Graph for Gemini. Next: Merchant Center, UCP and shopping visibility in Gemini. Or return to the Gemini for Marketers track. Related: How to rank in Gemini, Content formats that get cited in Perplexity, Get your brand into AI Overviews.
12Common questions
What content does Google say works best for generative AI search? Unique, non-commodity content that is helpful, reliable and people-first, such as first-hand experience and expert takes, organised with clear headings and supported by images and video.
Should I write separate pages for every fan-out query? No. Google says creating pages for every query variation mainly to manipulate AI responses violates its scaled content abuse policy and is ineffective.
Do I need to chunk content for AI? No. Google says there is no requirement to break content into tiny pieces, and there is no ideal page length.
Questions people ask
3 answersWhat content does Google say works best for generative AI search?
+
Unique, non-commodity content that is helpful, reliable and people-first, such as first-hand experience and expert takes, organised with clear headings and supported by images and video.
Should I write separate pages for every fan-out query?
+
No. Google says creating pages for every query variation mainly to manipulate AI responses violates its scaled content abuse policy and is ineffective.
Do I need to chunk content for AI?
+
No. Google says there is no requirement to break content into tiny pieces, and there is no ideal page length.
Sources
- Google Search Central, Optimizing your website for generative AI featuresdevelopers.google.com
- Google Search Central, Creating helpful, reliable, people-first contentdevelopers.google.com
- Google Blog, More ways to connect with the web from Search (15 Oct 2026)blog.google
- Google Search Central, AI features and your websitedevelopers.google.com
- All accessed 5 Nov 2026.