AI search does not answer a single question: it takes the buyer decision process apart and cross-checks evidence along the chain. If you have staffed only one layer (say, chasing generic queries), AI cannot find you when it drills deeper, and the citation goes to a competitor with fuller coverage. This lesson splits the decision chain into five layers L1–L5 and tells you what content each layer needs, and which to do first.
What the full-intent system is
Direct answer: Split the whole buyer decision chain, from “not knowing” to “placing an order”, into five layers L1–L5, with matching content for each. AI search cross-checks layer by layer along that chain, so covering only one layer gets you down-weighted for breaking the chain.
| Hierarchy | Buyer state | Content to staff it with | Keyword examples |
|---|---|---|---|
| L1 Awareness | Just realising the problem or category | Category explainers, what is X, use cases | what is flexible solar panel |
| L2 Solution | Looking for solution types | Solution comparisons, selection guides, type overviews | flexible vs rigid solar panel |
| L3 Query | Generic query, most mixed intent | Reviews, rankings, generic Q&A (most contested) | best solar panel supplier |
| L4 Decision | Comparing specs, certifications, price | Spec tables, certification notes, FAQ, MOQ / lead time | 100W solar panel specification |
| L5 Vendor | Verifying whether a specific factory can do it | Factory strength, production lines, case studies, qualifications | 100W flexible panel OEM China |
Why L3-only gets squeezed out
Direct answer: L3 generic queries are the most contested, and once AI has an answer it keeps verifying downwards into specifications and suppliers. With only L3 covered, the chain breaks at L4 and L5, and AI cites a competitor with more complete coverage instead.
Many teams treat GEO as “ranking for a handful of generic query sentences”. That is old-school SEO thinking applied to AI search. A buyer decision inside AI is a chain: what is it (L1), how to choose (L2), what is recommended (L3), checking specifications (L4), and finally confirming who can actually make it (L5). AI looks for evidence layer by layer along that chain — with only L3 covered, there is nothing to catch the later layers, and the citation goes to a rival whose chain is complete.
Rolling out the five layers and priorities
Direct answer: Fill L4 and L5 first (closest to the order, highest conversion), then build L1 and L2 context with a small amount of content; L3 is the most contested, so leave it last and enter through comparison and review content.
- Do L4, the decision layer, first: Spec tables, certifications (CE / UL / RoHS), MOQ, lead time, warranty — the things B2B buyers actually check.
- Then L5, the vendor layer: Real production-line photos, capacity, client case studies, qualification documents — answering “can this factory actually do it”.
- Fill in L1 and L2 context: Two or three explainer and selection pieces per category is enough to let AI understand where you sit in the category.
- Leave L3 for last: Enter through comparison tables, reviews and FAQ rather than fighting head-on for generic head terms.
How to find your content gaps
Direct answer: Export your Google Search Console queries, group them by the five layers, and find the layers that have queries but no matching content; then search your core term and see which layers the top-ranking competitors cover.
- Step 1 · Group your existing queries: Group your GSC queries by L1–L5 and see which layers get impressions but have no landing page.
- Step 2 · Check competitor coverage: Search your core term and check, page by page, which layers the top results cover; whatever is missing is your opportunity.
- Step 3 · Cross-validate: Cross-check against the three buying stages in Stage 2, Lesson 5 (awareness / consideration / decision) and prioritise the layers both approaches point to.
⚠️ Note: intent layering is an industry analysis framework, not an official Google classification; the number of layers and their names are not consistent across methodologies. Roll it out against your own content gaps and do not manufacture empty pages just to fill five layers — what Google officially emphasises is always “helpful, people-first content”.
FAQ
- Is the full-intent system an official Google concept?
- No. It is an industry analysis framework for taking the AI search decision chain apart; several agencies use it and neither the number of layers nor their names are standardised. What Google officially emphasises is “helpful, people-first content”. Roll it out against your own content gaps and do not manufacture empty pages just to fill five layers.
- With a small content team, do we need all five layers?
- You do not have to fill them all at once. Prioritise L4 (decision) and L5 (vendor) — those two are closest to the order and convert best; a small amount of L1 and L2 content is enough to establish category context; L3 generic queries are the most contested, so leave them last and enter through comparison and review content.
- How do I tell which layer my content is missing?
- Export your Google Search Console queries, group them by the five layers and see which layer has queries but no matching content; then search your core term and see which layers the top-ranking competitors cover. Put the two side by side and the gaps appear.
Keep Learning (by Track)
← Previous · Stage 4, Lesson 3
AI Visibility Monitoring and Data Review
🏁 You have completed all 16 lessons