How SolCrys Works
SolCrys editorial standards
These are the rules SolCrys uses to decide what is and isn't safe to publish about AEO, GEO, AI search, and Google AI features. We publish them because the AEO category has a credibility problem — too many vendors recommend tactics that Google has now explicitly said don't work, and buyers have no easy way to tell which vendors are intellectually honest. The standards have 5 mandatory DO's (non-commodity angle, user-satisfaction test, canonical consolidation, visible evidence, anti-hack check) and 7 named anti-patterns we will not recommend even when buyers ask for them. We will update this page as the AI search landscape evolves and our understanding sharpens. If you see SolCrys content that violates these standards, please tell us — we will fix it or retract it.
Updated 2026-05-16
Questions this guide answers
- How do you evaluate an AEO vendor's intellectual honesty?
- What AEO tactics does SolCrys refuse to recommend?
- How does SolCrys decide what content to publish?
- What does Google say about llms.txt and AI schema?
Why we publish this
The AEO category has a credibility problem. In the 24 months since AI Overviews launched, a substantial fraction of vendor marketing has converged on a set of tactics — `llms.txt`, AI-only schema, forced content chunking, FAQ-schema quotas, page factories for every long-tail variation — that have no documented support from the major AI engines. Some of those tactics have now been explicitly rejected by Google's own May 2026 generative AI optimization guide.
Buyers evaluating AEO platforms in 2026 are right to be skeptical. The easiest signal we know how to give them is the one on this page: SolCrys publishes the rules it uses to decide what is and isn't safe to publish. If you see something on solcrys.com that violates these standards, tell us — we will fix it or retract it. The standards are versioned and dated below, and we will update them as the AI search landscape evolves.
The 5 mandatory DO's
Every SolCrys public-facing asset (website article, LinkedIn long-form, lead magnet, sales playbook, ad copy, methodology page) must satisfy all five before it ships. If an author can't point to where each one lives in the draft, the asset isn't ready.
1. Non-commodity angle
Every piece must contain at least one of: a unique point of view that doesn't already exist on the first page of Google for the topic; a first-hand observation from SolCrys customer work (anonymized when required); an original framework or named methodology; or a verifiable data point from named primary research. If the draft can be reproduced by paraphrasing the top 3 SERP results, it is commodity content and gets killed or rewritten.
2. User-satisfaction test
After reading the asset, the target reader should be able to complete a real task, make a real decision, or diagnose a real problem. If the reader's takeaway is just 'now I know what AEO is in general,' the piece is too thin to ship.
3. Canonical consolidation plan
Before drafting a new page, the author must answer: is there already a SolCrys page covering this topic? If yes, the default is to update the existing canonical page, not create a new one. A new page is only justified when the user intent, audience, conversion path, or evidence requirements clearly differ from any existing asset. This rule exists because Google explicitly flags 'creating separate content for every possible variation of how people might search' as scaled content abuse spam — and because thin pages dilute our own authority.
4. Visible evidence and assets
Every asset's brief must specify what visual or structural evidence the piece needs: screenshots (for arguments about something visible), tables (for comparisons), charts of original SolCrys data (for quantitative claims), diagrams (for new frameworks), methodology disclosure (when referencing internal data). Text-only assets are allowed for short essays and FAQs, not for guides, methodologies, or vendor comparisons.
5. Anti-hack check
A draft fails the anti-hack check if it recommends, without explicit refutation, any of the 7 named anti-patterns below as a path to AI search visibility. Refuting these patterns in an essay is welcome and encouraged. Recommending them as Google-AI optimization is a hard fail.
The 7 named anti-patterns we will not recommend
These are the tactics that have circulated in the AEO category and that we have decided not to recommend, even when buyers ask. For each one, we name the reason in plain language.
| Anti-pattern | Why we don't recommend it |
|---|---|
| llms.txt or any AI-specific text/markup file as a Google AI lever | Google explicitly says in its May 2026 guide that no machine-readable AI files are required. OpenAI, Anthropic, and Perplexity have not committed to reading llms.txt either; when they want a new retrieval path they introduce a named crawler and document it. |
| AI-only schema.org markup | Google explicitly says there is no special schema for AI features. Schema remains useful as general SEO hygiene — but only when it matches visible content. |
| FAQ schema on pages without visible FAQs | Mismatched structured data is a Google policy issue and may be ignored or trigger structured-data penalties. |
| Forced content chunking for Google AI surfaces specifically | Google explicitly says there is no requirement to break content into tiny pieces. Structural clarity is good editorial practice for human readers; it is not a Google AI ranking factor. |
| One page per fan-out / PAA / long-tail variation | Google flags this as scaled content abuse spam. Consolidate into authoritative pillar pages instead. |
| Promising guaranteed AI citation, ranking, or visibility | Indefensible. Opens the buyer to procurement-level pushback when results vary by engine and time. |
| Inauthentic brand mentions (paid placements without disclosure, review-farm posts, AI-generated low-effort engagement) | Google explicitly flags this as not helpful. Brand-risk negative. Earned signals compound; manufactured ones don't. |
Engine-specificity rule
We do not publish generic 'AI engines do X' claims that don't hold across every named engine. Where a claim is engine-specific, we attribute it: 'ChatGPT extracts in passages,' 'Perplexity's RAG retrieval,' 'Google AI Mode in some modes.' This rule matters because different engines have meaningfully different retrieval architectures — Google AI Overviews and AI Mode are grounded in Google's Search ranking systems, while ChatGPT Search, Perplexity, Claude with web search, Gemini grounding, and retail RAG (Rufus, Sparky, ChatGPT Shopping) are true RAG architectures where structural clarity helps passage retrieval. Lumping them together produces bad advice.
Technical-honesty principles
Three rules that protect every SolCrys claim in enterprise procurement review.
- No 'studies show' without a citation. Either name the study or remove the claim.
- No 'AI engines now' without engine specificity. '[Named engine] now penalizes X, [date and source]' is the bar.
- No 'increases citation by N%' without a methodology section. Either disclose how the lift was measured (sample, period, control, statistical method) or strip the number.
The pre-publication checklist
Before any SolCrys public asset ships, the author runs an automated grep for the 7 anti-patterns plus 7 explicit human-judgment questions. The full checklist is published internally in our GEOResearch repository and applied to every release. If you would like to see the operational form, the evaluate-aeo-platform-methodology-checklist page is the public-facing version.
How to flag a violation
If you see SolCrys content — on this site, in LinkedIn long-form, in a sales deck, anywhere — that recommends one of the 7 anti-patterns, or fails one of the 5 DO's, we want to hear about it. Email the team or send a note via the contact form. We will fix the asset, retract the claim, or update the standards. The version and change-log below get updated when that happens.
Version and changelog
Current version: v1.0 — published 2026-05-16. Initial extraction from internal methodology documents following Google's May 2026 generative AI optimization guide. Future updates will be dated below.
FAQ
Do other AEO vendors publish editorial standards?
Most do not. A few publish methodology pages that describe their data pipeline; very few publish anti-patterns they refuse to recommend. We think the absence is itself a signal — if a vendor's content can recommend any tactic that brings in a deal, the vendor's advice is not load-bearing.
What if a customer specifically asks SolCrys to recommend llms.txt or FAQ-schema stuffing?
We say no, and we point to this page. The customer is free to implement those tactics on their own; we are not obligated to be the vendor that legitimizes them.
Are these standards versioned?
Yes. v1.0 is dated 2026-05-16. Every change gets a dated entry in the changelog and a version bump (minor for additions, major for breaking changes). We will not silently revise published positions.
Where can I see the internal version of this document?
The full internal SOP — including the operational pre-publication grep and the 7 human-judgment questions — lives in our GEOResearch repository under solcrys_content_system/. We share it with prospective enterprise customers under NDA. This page is the public summary.
How often do you update the standards?
Whenever a major AI engine publishes new official guidance (the May 2026 Google guide is the most recent trigger), whenever we discover a new anti-pattern in the wild, or quarterly at minimum. Updates are versioned and dated.
Related guides
Strategy & Positioning
Why llms.txt Is Not a Strategy
llms.txt is a proposed standard for AI-friendly content delivery, but it is neither widely adopted by major AI engines nor a substitute for AEO fundamentals. This essay explains what llms.txt does, what it does not, and why brands should focus on the unsexy basics.
Buyer Guides
Evaluate an AEO Platform's Data Methodology
Six questions every buyer should send to every AEO platform - including us - before signing. We designed SolCrys to answer all six; here's how, and what to listen for from anyone you're evaluating.
AEO Fundamentals
Answer Engine Optimization
Learn what Answer Engine Optimization is, why it matters, and how brands can structure content to be cited accurately by AI answer engines.
Free AI visibility audit
Find out where your brand is missing, miscited, or misrepresented.
SolCrys maps high-intent prompts to mentions, citations, answer accuracy, and content gaps so your team can prioritize the next pages to ship.