Google's spam policies now cover attempts to manipulate generative AI responses.
Source: Google Search spam policies, clarified May 2026.
Principles
The Clean AEO Standard
The Clean AEO Standard is six rules for improving how AI assistants describe a company without faking anything. We follow it on every account and publish it so any agency, in-house team or vendor can adopt it.
Six rules
What clean AEO means
- 01
Odds, not ranks
Every number comes with a sample size and a margin of error. AI answers change from run to run, so a single "position" is noise. Report how often, how sure, and how many runs.
- 02
Disclosed-only offsite work
No fake reviews, no sock-puppet accounts, no paid "organic" posts, no hidden prompts, and no "independent" comparison pages the client secretly controls. Every placement is logged with who posted it, where, and how it was disclosed.
- 03
No guarantees
Nobody controls what an AI assistant says. Sell a process and measured change against controls, never "#1 in ChatGPT".
- 04
Published prices and plain terms
Buyers should know what they'll pay and how to leave before the first call.
- 05
Client data stays the client's
Personal data is redacted before any model sees it. Each client gets a separate workspace with a retention date, and a data processing agreement is signed before anything changes hands.
- 06
A human signs off on everything published
AI can draft. A named person checks the facts against approved sources and approves every page, placement and report before it goes out.
Off the table
What we don't do
- Guarantee positions or outcomes
- Seed Reddit, Quora or forums with undisclosed posts, or buy aged accounts
- Write fake reviews or build "independent" comparison sites the client controls
- Use hidden text, prompt injection or "AI memory" tricks
- Run fleets of scrapers logged into client or personal accounts
- Sell llms.txt files or "AI-only" markup as a growth lever
For online stores
Extra rules for commerce
Selling products adds its own temptations. These apply to every Shopify engagement.
- 01
Real reviews only
No fake, bought or gated reviews. Incentivized reviews are disclosed. The FTC's 2024 rule on consumer reviews and testimonials bans fake and bought reviews outright.
- 02
Disclosed placements
No undisclosed paid placements in "best of" roundups. Affiliate and gifting relationships are disclosed wherever the placement appears.
- 03
The feed matches the page
Product feeds and markup always match what shoppers see: the same price, stock and claims. No hidden text or prompts aimed at AI shopping agents.
- 04
No fake shoppers
No sock-puppet Reddit, forum or community posts. A brand's team takes part openly, as itself.
Why now
Manipulation is now policed
Clean work used to be a preference. It's now a risk decision for anyone buying AEO help.
Bing's webmaster guidelines added a "Prompt Injection and AI Manipulation" section, with demotion or delisting as possible penalties.
Source: Bing Webmaster Guidelines, February 2026.
The FTC finalized a $750,000 civil penalty against a company over bot-run fake profiles and undisclosed employee reviews.
Source: FTC final order, TruHeight, July 2026.
Reddit began removing AI-optimization seeding at scale, catching about 25,000 spammy posts a day.
Source: The Next Web, July 2026.
Signatories
Who has adopted it
No signatories yet beyond Pop Sota. Be the first to adopt it.
Adopt
Adopt the Standard
Tell us who you are. We'll confirm the details with you before listing anyone, and you can stay unlisted.