# GEO–AEO–LLMO–SEO Operating Brief

Use this working document to turn an acronym-led request into a governed search and discovery program. Duplicate it for each market, product line, or materially different audience.

Version: ____________________  
Owner: ____________________  
Review date: ____________________  
Markets/languages: ____________________

---

## 1. Decision statement

### Business decision to improve

What decision should a buyer, user, employee, or agent make more successfully?

> ______________________________________________________________________

### Audience and journey stage

- Audience: _____________________________________________________________
- Journey stage: [ ] Discover  [ ] Compare  [ ] Validate  [ ] Buy/act  [ ] Support
- Current friction: ____________________________________________________
- Desired qualified outcome: ___________________________________________
- Explicit non-goals: __________________________________________________

### The four-field acronym decoder

| Field | Required answer |
|---|---|
| Target surface | Google Search / AI Overviews / AI Mode / ChatGPT Search / Perplexity / owned RAG / agent / other: __________ |
| Controllable intervention | Content / template / access / entity facts / independent corroboration / feed / retrieval / model / other: __________ |
| Observable evidence | Indexation / impression / retrieved source / mention / citation / referral / task success / other: __________ |
| Business outcome | Qualified visit / lead / sale / opportunity / resolved task / productivity / other: __________ |

**Scope test:** If any field is blank, do not approve the acronym as a deliverable.

---

## 2. Search-lens selection

Check each lens that has a proven role. One item can support several lenses.

| Lens | Use when | Evidence that it belongs in scope | Accountable owner |
|---|---|---|---|
| SEO | Search access, indexation, relevance, authority, experience, or organic measurement is the constraint |  |  |
| AEO | A correct answer is buried, ambiguous, unscoped, or hard to compare |  |  |
| GEO | Public generated responses are commercially relevant and source/brand visibility is not understood |  |  |
| LLMO | A named LLM, retrieval system, model, feed, or agent needs a controlled optimization/evaluation program |  |  |

### First proven bottleneck

[ ] Not crawled/rendered/indexed → SEO foundation  
[ ] Answer buried or unclear → AEO clarity  
[ ] Eligible but not used/mentioned/cited → GEO evidence  
[ ] Owned RAG/model/feed/agent → LLMO system  
[ ] Measurement cannot support a decision → analytics first

Priority and rationale:

> ______________________________________________________________________

---

## 3. Question-journey cohort

Use a stable cohort for before/after observation. Record the exact wording; do not replace it silently between runs.

| ID | Stage | Exact question/prompt | Market/language | Intended information need | Relevant owned page | Priority |
|---|---|---|---|---|---|---|
| Q01 | Discover |  |  |  |  | High / Med / Low |
| Q02 | Compare |  |  |  |  | High / Med / Low |
| Q03 | Constraint |  |  |  |  | High / Med / Low |
| Q04 | Evidence |  |  |  |  | High / Med / Low |
| Q05 | Next step |  |  |  |  | High / Med / Low |

Sampling controls:

- Platform/interface: _________________________________________________
- Account or logged-out state: ________________________________________
- Device/location controls: ___________________________________________
- Run frequency: ______________________________________________________
- Capture method and storage: _________________________________________
- Known personalization/volatility limitations: _______________________

---

## 4. Access and purpose register

Confirm current vendor documentation before changing a control. Access can support eligibility; it cannot guarantee retrieval, citation, recommendation, or training behavior.

| Operator/token | Intended purpose | Policy decision | robots/meta/WAF state | Log or test evidence | Owner | Last checked |
|---|---|---|---|---|---|---|
| Googlebot | Search discovery/indexation | Allow / Block / Limited |  |  |  |  |
| OAI-SearchBot | ChatGPT search surfacing | Allow / Block / Limited |  |  |  |  |
| GPTBot | Potential OpenAI model training | Allow / Block / Limited |  |  |  |  |
| ChatGPT-User | User-initiated fetch/actions | Allow / Block / Limited |  |  |  |  |
| PerplexityBot | Perplexity search surfacing | Allow / Block / Limited |  |  |  |  |
| Perplexity-User | User-initiated fetch | Allow / Block / Limited |  |  |  |  |
| Other: ______ |  | Allow / Block / Limited |  |  |  |  |

Technical checks:

[ ] `robots.txt` response and relevant groups inspected  
[ ] Meta robots and `X-Robots-Tag` inspected  
[ ] Authentication/cookie wall behavior inspected  
[ ] CDN/WAF rules and verified IP guidance inspected  
[ ] Server-side and rendered HTML compared  
[ ] Canonical, status, redirects, and snippet eligibility checked  
[ ] Server/CDN logs sampled where lawful and available  
[ ] Legal, privacy, licensing, and data owners recorded

---

## 5. Source-asset inventory

| Page/asset | Decision served | Unique evidence | Named expert | Method/scope/date visible | Primary sources | Extractable answer | Maintenance owner | Action |
|---|---|---|---|---|---|---|---|---|
|  |  |  |  | Yes / No |  | Yes / No |  | Keep / Improve / Create / Retire |
|  |  |  |  | Yes / No |  | Yes / No |  | Keep / Improve / Create / Retire |
|  |  |  |  | Yes / No |  | Yes / No |  | Keep / Improve / Create / Retire |

### Source-quality gate

For every priority asset:

[ ] Answers the primary question before long background  
[ ] Defines terms and comparison criteria  
[ ] Places conditions, markets, units, and dates beside the claim  
[ ] Uses inspectable first-party evidence or primary references  
[ ] Names authors/experts and their relevant experience  
[ ] Discloses method, sample, assumptions, and limitations where needed  
[ ] Includes a useful diagram, example, calculation, or procedure where it improves understanding  
[ ] Offers a helpful next step after the short answer  
[ ] Has a review trigger and maintenance owner  
[ ] Avoids unsupported ranking factors, fabricated statistics, and guaranteed citations

---

## 6. Entity and fact register

| Fact/entity | Approved value | Scope/market | Source of truth | Owned URLs checked | External profiles checked | Conflict found | Owner |
|---|---|---|---|---|---|---|---|
| Brand/legal name |  |  |  |  |  |  |  |
| Product/service name |  |  |  |  |  |  |  |
| Availability/location |  |  |  |  |  |  |  |
| Price/plan |  |  |  |  |  |  |  |
| Policy/guarantee |  |  |  |  |  |  |  |
| Expert/author |  |  |  |  |  |  |  |
| Other: ______ |  |  |  |  |  |  |  |

Never use structured data to assert a fact that is absent from or contradicted by the visible page.

---

## 7. Visibility and business baseline

Keep every measurement layer separate.

| Layer | Definition for this program | Data source | Baseline/window | Target or decision threshold | Limitation |
|---|---|---|---|---|---|
| Eligible |  |  |  |  |  |
| Retrieved/used |  |  |  |  |  |
| Mentioned |  |  |  |  |  |
| Cited |  |  |  |  |  |
| Visited |  |  |  |  |  |
| Converted |  |  |  |  |  |

### Observation record

| Run date | Prompt ID | Platform | Brand mentioned? | Linked citation? | Cited URL/domain | Answer accurate? | Competitors/sources | Evidence location |
|---|---|---|---|---|---|---|---|---|
|  |  |  | Yes / No | Yes / No |  | Yes / Partial / No |  |  |
|  |  |  | Yes / No | Yes / No |  | Yes / Partial / No |  |  |

### Measurement cautions

- Do not treat one generated answer as a stable rank position.
- Do not count a mention as a citation.
- Do not count platform impressions twice when a generative report is included in broader search totals.
- Do not infer incremental revenue from correlation alone.
- Record prompts, environment, sample size, missing data, volatility, and method changes.

---

## 8. Experiment card

Experiment ID/name: ____________________________________________________

| Field | Definition |
|---|---|
| Evidence-backed problem |  |
| Hypothesis |  |
| Target question IDs/surfaces |  |
| Intervention |  |
| Control or comparison |  |
| Primary observable |  |
| Guardrail metric |  |
| Business signal |  |
| Start/re-test dates |  |
| Cost/owner |  |
| Confounders |  |
| Decision rule | Keep / Revise / Stop when: __________________ |

Result and interpretation:

> ______________________________________________________________________

Do not change content, outreach, technical access, and sampling method at the same time if the goal is to learn which intervention mattered.

---

## 9. Shared 90-day backlog

| Period | Work item | Lens(es) | Evidence/problem | Deliverable | Acceptance evidence | Owner | Dependency | Status |
|---|---|---|---|---|---|---|---|---|
| Days 1–30 | Define surfaces and prompt cohort | GEO/LLMO |  |  |  |  |  |  |
| Days 1–30 | Technical/access baseline | SEO/GEO |  |  |  |  |  |  |
| Days 1–30 | Visibility/business baseline | All |  |  |  |  |  |  |
| Days 31–60 | Fix first proven bottleneck |  |  |  |  |  |  |  |
| Days 31–60 | Improve/create source asset | AEO/GEO/SEO |  |  |  |  |  |  |
| Days 61–90 | Re-test stable cohort | GEO/LLMO |  |  |  |  |  |  |
| Days 61–90 | Evaluate qualified outcomes | All |  |  |  |  |  |  |
| Days 61–90 | Keep/revise/stop decision | All |  |  |  |  |  |  |

RACI/decision rights:

| Decision | Accountable | Responsible | Consulted | Informed | Response time |
|---|---|---|---|---|---|
| Search/crawler access |  |  |  |  |  |
| Training permission |  |  |  |  |  |
| Editorial accuracy |  |  |  |  |  |
| Technical release |  |  |  |  |  |
| External claim/PR |  |  |  |  |  |
| Measurement method |  |  |  |  |  |

---

## 10. Vendor-scope audit

Require a written answer for every item.

[ ] Named target platforms/interfaces and markets  
[ ] Exact deliverables and controllable changes  
[ ] Prompt/question sampling method  
[ ] Baseline, denominator, time window, and calculation for every uplift claim  
[ ] Separation of mention, citation, referral, and revenue  
[ ] Separation of search crawling, training, and user-initiated fetching  
[ ] No guaranteed rank, citation, recommendation, or revenue  
[ ] No proprietary score presented as a platform metric  
[ ] No “special AI schema” or `llms.txt` claim presented as a Google ranking factor  
[ ] Source and content quality standard  
[ ] Data ownership, privacy, retention, and access terms  
[ ] Re-test protocol, limitations, and stop rule  
[ ] One shared backlog with existing SEO, editorial, engineering, PR, and analytics owners

Unresolved commercial or technical risks:

> ______________________________________________________________________

---

## 11. Review decision

Review date: ____________________

### Evidence summary

- What improved: _______________________________________________________
- What did not improve: _______________________________________________
- What remains unobservable: __________________________________________
- What changed in the environment/method: _____________________________
- Qualified business evidence: ________________________________________

### Decision

[ ] Scale the intervention  
[ ] Continue collecting evidence  
[ ] Revise the intervention  
[ ] Fix a different upstream bottleneck  
[ ] Stop the work  
[ ] Reclassify the scope (SEO / AEO / GEO / LLMO / product / data / analytics)

Decision owner and rationale:

> ______________________________________________________________________

Next review trigger/date: ______________________________________________

---

This brief is an operational template, not a promise of ranking, inclusion, citation, recommendation, or revenue. Validate crawler behavior and platform reporting against current official documentation before making policy or investment decisions.