manifoldmcp/marketing-skills / plugins/manifold-ai-search/skills/ai-search
ai-search
VerifiedWhen the user wants AI visibility, whether ChatGPT, Claude, Gemini, Perplexity and Google AI Overviews mention or cite the brand, which sources they cite, and how to get cited (GEO, AEO, LLM SEO). Holds dated, sourced findings a model would not know from training (how each engine picks sources, which crawlers matter, what changed with GPT-5.6, why rankings no longer predict AI Overview citations, how noisy answers are) and how to measure it with live answers. Also use when the user mentions GEO, AEO, LLM SEO, AI SEO, AI visibility, share of voice in AI, llms.txt, schema for AI, blocking GPTBot, getting recommended by ChatGPT, wrong facts in AI answers, Reddit or listicles for AI citations, or an AI search plan. Classic Google rankings go to seo, a technical crawl to technical-seo.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 10.9 KB |
| references/citations.md | markdown | 6.5 KB |
| references/engines.md | markdown | 6.2 KB |
| references/field-notes.md | markdown | 4.8 KB |
| references/google-ai-overviews.md | markdown | 4.9 KB |
| references/measurement.md | markdown | 5.8 KB |
| references/site-access.md | markdown | 4.2 KB |
| references/strategy.md | markdown | 3.1 KB |
| references/wrong-answers.md | markdown | 2.9 KB |
references/strategy.md
View on GitHub# An AI search plan
When the user wants a plan rather than a check: which buyer questions to win, where the brand stands on each engine, and which work closes the gap fastest, over 90 days.
## Order of work
The order follows what blocks what. Skip a step only when the baseline shows it is not a problem.
1. **Access.** Any `blocks_citations` failure from `aeo_get_site_readiness` comes first, because nothing else works while engines cannot read the site ([site access](site-access.md)).
2. **Wrong facts.** An answer that gets pricing or capability wrong costs sales now ([wrong answers](wrong-answers.md)).
3. **Third-party sources.** The cited pages that name competitors and not the user, ranked by engines and prompts citing them ([citations](citations.md)). This is usually the bulk of the work and the biggest lever, because mentions follow being talked about elsewhere.
4. **Owned pages.** Pages that answer the buyer questions directly and are kept current: pricing, comparison (honest, not self-ranking), use-case and docs pages. Substantive updates to existing pages beat new pages ([citations](citations.md#freshness)).
5. **Google.** Where AI Overviews or AI Mode matter, measure them per keyword: ranking helps but no longer predicts citation ([Google's AI answers](google-ai-overviews.md)).
Choose per engine. A B2B brand losing on ChatGPT needs off-site mentions and editorial coverage; a local business losing on Gemini needs its own site and listings; a shop on ChatGPT needs its product feed.
## Baseline
A fixed prompt set of ten to forty buyer prompts on the engines the buyers use, scored as [measurement](measurement.md) says, with the margin of error beside each share. About 364 credits with Manifold for ten prompts on five engines plus the prompt research; 270 for each re-measure. Readiness is free.
## The 90 days
| Period | Typical work | KPI the tools re-measure |
|---|---|---|
| Days 1 to 30 | Access fixed, wrong facts corrected, first 15 cited sources contacted, owned answer pages updated | Readiness checks passing; wrong claims gone on re-run |
| Days 31 to 60 | List and editorial inclusions, Reddit and YouTube participation where those engines cite them, new or rebuilt answer pages | Cited sources now naming the user |
| Days 61 to 90 | Second round of sources, re-measure on the same prompt set | Mention share and citation share per engine, against the margin |
Give each period an owner, and three first actions for this week.
## Setting targets honestly
- Promise inputs, not positions: a readable site, correct facts, presence on the sources engines cite. No one can promise a place in an answer, and lists rarely repeat ([measurement](measurement.md#noise)).
- Set KPI targets larger than the margin of error on the set's size. On 50 answers a ten-point move may be noise.
- Engines that search live can move in weeks; answers from training data move on model releases. Say which is which when setting the 30-day target, and say that no study has measured how long a change takes to show.
- A model release can reshape an engine overnight (GPT-5.6 cut listicle and Reddit citations in August 2026). Re-check the plan's assumptions after any major release.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack." | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |