aeko-brand-competitor-analysis

AEKO Brand Competitor Analysis

Produces a competitor-positioning analysis at the brand level (not product level — for that use /aeko-product-competitor-analysis). Output: a clear picture of what the competitor stands for publicly, where they win AI answers, and one concrete action the user could take.

Marketer-facing output contract

Frame this as "why AI may mention this competitor instead of us." Start with public positioning, AI visibility gap, and one practical next move. Keep source caveats clear; never imply AEKO has measured a signal when it is inferred from WebSearch/WebFetch.

Language: mirror the user's chat language for user-facing steps, summaries, questions, and next actions. Keep slash commands, IDs, file paths, channel slugs, schema keys, and tool names in English/ASCII.

Inputs

• domain-id (optional) — UUID. Missing → aeko_list_domains pick-list.
• competitor (required) — Competitor name OR root domain (e.g. "필리" or "pilly.co.kr"). Passed as $1 (if no domain arg) or $2.

Step 1 — Resolve user's domain + brand context

1. Parse $1 / $2. If first positional is clearly a UUID → domain_id, second is competitor. Else treat the positional as competitor and resolve domain via aeko_list_domains.
2. aeko_get_domain_info(domain_id) + aeko_get_brand_kit(domain_id) to ground the "vs us" comparison.

If competitor is a name (not URL), run a quick WebSearch for "<competitor>" official site to resolve to a root domain. Confirm with the user if ambiguous.

Step 2 — Gather public signals on the competitor

Collect in parallel where possible:

1. Root page crawl — WebFetch(<competitor_root>). Extract: tagline, hero messaging, top-nav categories, whether llms.txt / structured data is present.
2. Wikipedia / Wikidata entity check — WebSearch("<competitor> site:wikipedia.org") and WebSearch("<competitor> site:wikidata.org"). If results → the competitor has AI-knowledge-graph recognition, which is a load-bearing signal.
3. Recent news — WebSearch("<competitor> news 2026") (adjust year if needed). Surface any fundraising, product launches, brand refresh.
4. Press / partnerships — WebSearch("<competitor> partnership OR acquired OR launch") — optional if prose asks for depth.

Record the sources. Do NOT fabricate; if a search returns empty, note that.

Step 3 — Cross-reference AEKO citation data

This is the AEKO-grounded layer — skip this step and you're just a WebSearch wrapper.

1. aeko_get_visibility_summary(domain_id, scope="cited_sources") — surfaces pages from the user's domain AI engines cite.
2. For each of the user's top 5-10 tracked prompts (derive from brand_kit.brand_keywords[] + aeko_search_research_prompts(scope=..., country=...) if no tracked set yet):
   • Call aeko_get_tracked_prompt(prompt_id, window="30d") for cited-source analysis.
   • Count how often the competitor's brand name appears in responses[].mentions.
   • Count how often the competitor's root domain appears in responses[].citations[].domain.
3. Build a comparison matrix:

| Prompt | Our mentions | Competitor mentions | Our citations | Competitor citations |
|--------|--------------|---------------------|---------------|----------------------|
| ...    | ...          | ...                 | ...           | ...                  |

Rank prompts by competitor_mentions + competitor_citations - our_mentions - our_citations. Top rows = prompts where the competitor is winning and the user isn't.

Step 4 — Write the analysis

Compose a markdown report:

# Competitor Analysis: <competitor>
**Against:** <user's brand> (`<domain>`)
**Generated:** <ISO date>

## Public positioning

- **Tagline:** <from hero crawl, or "none found">
- **Top categories:** <from top-nav>
- **Wikipedia entity:** <yes with URL | no>
- **Wikidata entity:** <yes with URL | no>
- **Recent news:** <1-3 headlines, each with date and URL>

## AI visibility footprint vs <user's brand>

(Embed the comparison matrix from Step 3.)

**Interpretation:**
<one paragraph on what the matrix shows — e.g. "The competitor dominates
awareness-stage informational prompts in KR; we're stronger at
recommendation-stage prompts where brand reputation matters. The gap is
most pronounced in the top-3 prompts, where our brand is absent from
AI answers entirely.">

## What the competitor does that we don't

(From Step 2 signals that translate to AEO leverage:)
- Wikipedia entity → AI models have richer embeddings for them
- llms.txt on root → declared AI-readability
- Structured data present → Product / FAQPage / Organization schemas
- Recent news coverage → fresh signal for news-aware AI engines
- **Content frameworks** → where the competitor's *cited pages* win on substance, name it in the plugin's
  AEO vocabulary (BLUF / PREP / Informational Gain / E-E-A-T — see
  `skills/aeko-create-content/references/aeo-frameworks.md`) so the gap maps to a fix the executor skills apply.

For each present for competitor but absent for user → flag.

## Recommended action

<one concrete command the user could run to start closing the gap>

Step 5 — Save + summary

Write to ./aeko-artifacts/<domain_id>/competitor-analyses/<competitor-slug>-<YYYYMMDD>.md.

User-facing summary:

✔ Competitor analysis saved: <path>
  Competitor: <competitor>
  Biggest gap: <top row from Step 3 matrix>
  Next: <recommended action command>

Error paths

• Competitor name ambiguous + WebSearch returns multiple candidates → ask user to pick or paste the root URL directly.
• All cross-reference calls fail → still produce the Step 2 / Step 4 public-signals portion; note AEKO cited-source data missing.
• No tracked prompts + research-prompt fallback returns empty → note the user's domain doesn't have any tracked-prompt coverage yet; suggest /aeko-find-prompts-to-track as a prerequisite.

What this skill never does

• Never compares at the product level — that's /aeko-product-competitor-analysis.
• Never fabricates Wikipedia / news findings.
• Never posts anything externally.
• Never mutates tracking state.