AEO Visibility Audit — does AI actually recommend your business?

Find out whether ChatGPT, Perplexity, and Google's AI recommend your business — or your competitors — then get the exact sources you need to fix it.

Drop in one URL. This Actor figures out your brand and what you do, generates real buyer-intent questions for your city, asks them across the major AI answer engines, and returns a single structured JSON: your visibility score, who's getting recommended instead of you, the real Google AI Overview, the exact pages AI cites, and a prioritized action plan. Built for SEO/GEO agencies, local businesses, and anyone selling AEO services.

---

Why this one is different

Most "AI visibility" Actors send a prompt to a model and count mentions. This one adds the three things those tools leave out:

1. The REAL Google AI Overview. We pull the actual AI Overview from the live Google SERP (DataForSEO load_async_ai_overview) plus the local pack. Almost every competitor fakes this by querying the Gemini model and calling it "Google AI" — and several say so in their own docs. If the buyer's customers see a Google AI Overview, you should measure the real thing.
2. Real AI search volume (opt-in). Enable includeAiSearchVolume and each tracked prompt is enriched with DataForSEO AI Keyword Data — actual LLM prompt demand, not a made-up estimate. Straight talk on coverage: this is genuinely useful for national and brand-level prompts, but hyperlocal city-level phrasing mostly has no measured demand yet — in one 12-keyword Toronto sample, only one returned a volume. We return null rather than inventing a number, and it's off by default so you don't pay for empty fields on a local audit.
3. Brand → citation attribution. Not just "Competitor X was mentioned" and "these domains were cited" as two separate lists — we connect them: "Competitor X was named, and that mention was attributed to clutch.co and semrush.com." That's the causal map that tells you exactly which placement to go earn.

Plus: location-specific prompts (down to the city), a ranked action plan that names the competitor brands behind each source, and a shared cache that auto-refreshes on the 1st and 15th of each month so repeat lookups of the same URL are near-free.

---

Us vs. the other guys

	AEO Visibility Audit (this Actor)	LLM Visibility Tracker (khadinakbar)	LLM Visibility Monitor (constructive_calm)
Engines	ChatGPT, Perplexity, Google/Gemini	ChatGPT, Claude, Perplexity, Gemini	+ Claude, Grok
Real Google AI Overview (live SERP)	✅ Yes — actual AI Overview + local pack	❌ Gemini model as proxy	❌ Gemini model as proxy
Real AI search volume per prompt	✅ Opt-in (DataForSEO AI Keyword Data)	❌	❌
Never billed for a check that didn't search	✅	❌	❌
Brand → source attribution	✅ Yes — which sources produced each competitor mention	❌ (cited domains only)	❌ (share-of-voice only)
Auto-generated local buyer-intent prompts	✅ City-level	⚠️ Generic	✅ Category-level
Ranked action plan (names competitor brands)	✅	❌	⚠️ Gaps list
Shared cache + scheduled refresh	✅ 1st & 15th, repeat ≈ free	❌	❌
Output	One clean structured JSON	Per-result rows	Dataset + report files
Pricing (12-prompt audit)	$0.72 Quick · $2.65 Standard	~$4.32	~$6.90 (Quick preset $3.76)
Per-check price (metered)	$0.06	$0.09	$0.15 + $1.50 fixed fees

Reading the price right: competitors quote a per-check rate (e.g. "$150 / 1,000" = $0.15 for one prompt to one engine). A real audit is dozens of checks, so their full audits land at $3.76–$13.51 (their own published presets). We're the cheapest per check and bundle the real AI Overview and brand→source attribution they don't.

---

Two ways to pay

Mode 1 — Preset (one URL = one predictable price). Every preset is just a fixed number of billable checks, so the price is exact, not an estimate:

Preset	What runs	Price
Quick (default)	4 prompts × 3 engines, fast models, no AI Overview	$0.72
Standard	12 prompts × 3 engines + real AI Overview + attribution	$2.65
Deep	20 prompts × 3 engines, premium models + AI Overview	$4.09

Mode 2 — Metered (pay per check). Set preset: "custom" and only pay for what you ask for — ideal for power users and big or tiny audits:

Event	Price
Audit started	$0.00005
LLM visibility check (1 prompt × 1 engine)	$0.06
Real Google AI Overview module (per run)	$0.49
Brand → source attribution	Included free

You are never charged for a check that didn't search the web. If an engine answers from training data instead of running a live search, that check returns no citations — so we don't bill it. The count surfaces as summary.checks_without_web_search.

Cache hits incur only the start charge — no LLM cost for a repeat lookup inside the refresh window.

---

Input

Field	Description
url (required)	Homepage to audit. Brand + industry are auto-detected.
preset	quick (default), standard, deep, or custom (metered).
city, province, country	Localizes the generated prompts and the AI Overview SERP.
engines	Any of chat_gpt, perplexity, gemini.
maxPrompts	Prompts per engine (custom mode; presets set this).
includeAiOverview / serpMaxKeywords	Real Google AI Overview module (custom mode).
includeAiSearchVolume	Real LLM prompt-demand lookup (default off — see note below).
fastModels	Custom mode: use cheaper mini models.
cache / forceRefresh	Shared cache with 1st/15th auto-refresh.
dataforseoLogin / dataforseoPassword	Your DataForSEO credentials (owner-configured as Actor secrets for bundled runs).

Output (single structured JSON)

{
  "business": { "url": "...", "brand": "Flying Pigs Marketing", "descriptor": "marketing agency", "city": "Barrie" },
  "summary": { "prompts_run": 12, "engine_runs": 36, "mentioned": 3, "visibility_pct": 8,
               "ai_overview_appeared": 6, "ai_overview_you_cited": 0,
               "checks_without_web_search": 1, "dataforseo_cost_usd": 1.55 },
  "coverage": { "preset": "standard", "model_tier": "standard",
                "engines_checked": ["chatgpt", "perplexity", "google_ai"], "engines_not_checked": [],
                "ai_overview": "checked (6 keywords)",
                "ai_search_volume": "not run — includeAiSearchVolume was false",
                "citation_graph": "full", "upgrade": null },
  "by_engine": [ { "engine": "chatgpt", "model": "gpt-4o", "visibility_pct": 8 } ],
  "you_vs_them": { "yours": 2, "others": 71, "yours_pct": 3 },
  "google_ai_overview": { "source": "Google SERP ai_overview (real)", "appeared": 6, "you_cited": 0, "results": [ ] },
  "prompts": [ { "text": "best marketing agency in Barrie", "ai_search_volume": 210, "visibility_pct": 0,
                 "engines": [ { "engine": "chatgpt", "mentioned": false, "citations": ["clutch.co"],
                                "named_brands": [ { "brand": "Tactycs", "sources": ["tactycs.io"] } ] } ] } ],
  "top_sources": [ { "domain": "clutch.co", "times_cited": 9 } ],
  "competitor_attribution": [ { "brand": "Tactycs", "times_named": 5,
                               "attributed_sources": [ { "domain": "clutch.co", "count": 4 } ] } ],
  "action_plan": [ { "priority": 5, "title": "Earn placement / get cited on clutch.co",
                     "type": "directory", "competitors_present": ["Tactycs", "Thrive"] } ]
}

Methodology & honesty

LLM rows come from the engine APIs with live web search — a repeatable, labeled proxy that differs from a logged-in consumer app and is non-deterministic between runs. The google_ai_overview block is the real Google AI Overview scraped from the SERP. Every result is labeled by engine and model so you always know what produced it. This is a directional visibility audit for benchmarking and prioritization — not a traffic-analytics measurement.

---

Built by Scott Holmes — Pinnacle Tech Projects, Barrie, ON.