Google AI Overview Disruption Monitor for SEO Keywords

Triage your SEO keyword portfolio against Google's AI answers. For every keyword you get one decision row: whether an AI Overview is showing, whether AI Mode answers the query, who the AI cites, the classic search volume, and a single 0–100 AI disruption score that tells you which keywords to defend, monitor, or deprioritize.

Built for SEO leads, publishers, and e-commerce content strategists who need to know where AI is cannibalizing organic clicks — not another raw SERP dump. This measures click cannibalization per keyword, not search volume trends.

What does the AI Overview Disruption Monitor do?

Give it a list of keywords. For each one, the monitor checks Google's AI surfaces — the AI Overview block above the organic results and the conversational AI Mode answer — joins in the keyword's monthly search volume, and computes a disruption score plus a triage class. One keyword in, one scored row out.

It answers one question at scale: how much is Google's AI intercepting the clicks for this query, and how urgently should I act?

Why track AI Overview disruption

• One number per keyword. A disruption_score from 0 to 100 collapses AI Overview presence, format richness, AI Mode presence, search volume, and citation concentration into a single comparable metric.
• Ready-to-act triage. Every row is classed defend, monitor, or deprioritize — filter the dataset and hand your team a work list, no spreadsheet math required.
• Both AI surfaces, one row. AI Overview and AI Mode are checked and merged, so you see the full AI footprint on a query, not just one block.
• Citation intelligence. See exactly which domains the AI cites and how concentrated those citations are (a winner-take-all query scores differently from a broad one).
• Volume-weighted. A disrupted head term and a disrupted long-tail are not the same emergency — the score reflects exposure.
• Pay per keyword, no seat. Flat $0.35 per delivered row; repeat keywords within 24 hours are served from cache and cost nothing.

How to score AI Overview disruption for your keywords

1. Open the Actor and paste your keyword portfolio into the Keywords field (one phrase per line). Keywords are trimmed, lowercased, whitespace-collapsed, and de-duplicated, so two variants of the same phrase count as one row.
2. Run the Actor. It partitions cache hits first, enriches volume in a single batch call, then checks the AI surfaces for each new keyword.
3. Open the Overview dataset view for the at-a-glance list: keyword, triage class, disruption score, search volume, and AI Overview presence.
4. Switch to the Defend view to see only the highest-risk keywords (triage defend), ordered by disruption score — that is your priority defense list.
5. Export to JSON, CSV, or Excel, or pull the results from the API and feed them into your own SEO dashboard. Re-run every ~7 days to keep the picture current.

Minimal input:

{
  "keywords": ["best crm", "ai seo tool", "how to invoice a client"]
}

What data you get per keyword

One dataset row = one keyword. Every row always carries the same 16 keys (nothing missing, no raw provider payload leaks through). Example row:

{
  "keyword": "best crm",
  "status": "ok",
  "ai_overview_present": true,
  "ai_overview_format": "text_with_citations",
  "aio_cited_domains": ["hubspot.com", "salesforce.com", "zoho.com"],
  "ai_mode_answer_present": true,
  "ai_mode_answer_length": 1180,
  "ai_mode_cited_domains": ["hubspot.com", "forbes.com"],
  "ai_mode_has_links": true,
  "search_volume": 135000,
  "citation_concentration": 0.28,
  "disruption_score": 84,
  "triage_class": "defend",
  "source": "primary",
  "checked_at": "2026-07-23T09:14:02.371Z",
  "cache_hit": false
}

Field	Type	What it tells you
keyword	string	The normalized query (echo of input)
status	"ok" | "not_found"	not_found = no AI SERP data for the phrase — still a delivered answer
ai_overview_present	boolean	Is a Google AI Overview shown for this keyword?
ai_overview_format	string | null	Block richness: text, text_with_citations, or text_with_videos; null when absent
aio_cited_domains	string[]	Domains the AI Overview cites (processed, deduplicated)
ai_mode_answer_present	boolean | null	Does AI Mode answer the query? (null on the fallback path)
ai_mode_answer_length	integer | null	Length of the AI Mode answer in characters
ai_mode_cited_domains	string[] | null	Domains AI Mode cites
ai_mode_has_links	boolean | null	Does the AI Mode answer contain outbound links?
search_volume	integer | null	Classic monthly search volume; null when unavailable
citation_concentration	float 0–1	How concentrated the AI citations are across domains (HHI)
disruption_score	integer 0–100	The headline metric — see below
triage_class	"defend" | "monitor" | "deprioritize"	Decision bucket from the score
source	"primary" | "fallback"	Which data path filled the row
checked_at	ISO 8601	When this observation was taken
cache_hit	boolean	true when the row was served free from the 24h cache
tracked_domain_cited	boolean | null	true when any of your trackedDomains (or a subdomain) is cited in the AI Overview or AI Mode answer; null when you did not pass trackedDomains

AI disruption score and triage explained

The disruption_score (0–100) sums five weighted components:

• AI Overview presence — the single biggest click cannibalizer when an AI block sits above the organic results.
• AI Overview format — a richer block (citations, videos) holds the user longer than plain text.
• AI Mode presence — a second AI surface intercepting the same query.
• Search volume — log-scaled exposure, so head terms outweigh long-tail at equal disruption.
• Citation concentration — a winner-take-all query where the AI cites a handful of domains is scored higher than one that cites broadly.

Triage thresholds are fixed and inclusive: score ≥ 70 → defend, 40–69 → monitor, ≤ 39 → deprioritize. An unknown search volume never inflates a score.

Input parameters

Field	Required	Notes
keywords	yes	Your SEO keyword portfolio (min 1). Each phrase yields one scored row. Normalized and de-duplicated before scoring. Results use google.com in English.
trackedDomains	no	Up to 50 of your domains (bare hosts — protocol and www. are stripped). Each row then reports tracked_domain_cited: whether the AI answer cites your domain or any of its subdomains.

This actor is intentionally stateless in v1 — no geo, language, or location inputs. All keywords are checked against google.com in English. You can also run it programmatically via the Apify API and through the Apify MCP server (mcp.apify.com), so AI assistants can trigger it from your keyword list.

How much does the Google AI Overview Disruption Monitor cost?

Billing uses Apify's pay-per-event model. You pay per delivered row, not per input keyword and not per upstream call.

Event	When	Price
Run start	Once per run	$0.05 (flat, one-time)
Keyword scored	Per delivered row	$0.35 per keyword

The $0.35 per-keyword price is flat across every Apify plan tier. You are never charged for:

• Cache hits — a keyword scored in the last 24 hours is served from cache and bills nothing beyond the one-time run start.
• Failures — a keyword whose data could not be fetched is skipped and never billed.

A not_found result is a delivered answer and is billed as one keyword — knowing a query has no AI SERP data is a result. On FREE accounts the run stops before the first paid call once your plan limit is reached, so you are never charged beyond your balance.

Google AI Overview Disruption Monitor vs manual SERP checks

	This actor	Manual SERP checks	Rank trackers
Covers AI Overview	Yes	Yes, one query at a time	Rarely
Covers AI Mode	Yes	Manual, tedious	No
Volume-weighted score	Yes, 0–100	No	No
Ready-to-act triage	defend / monitor / deprioritize	No	No
Citation concentration	Yes (HHI)	No	No
Cost model	$0.35 per keyword, cache free	Your time	Monthly seat
Scales to a full portfolio	Yes, batched	No	Yes, but no AI signal

How the monitor stays reliable

• Cache (24h): identical keywords within 24 hours are served from cache — free, with the original observation time preserved.
• Deterministic scoring: the same inputs always produce the same score and class. The scoring version is part of the cache key, so a formula change cleanly invalidates old entries.
• Fallback path: if the primary source fails, the row is retried through a secondary source (source: "fallback"). On that path the AI Mode fields degrade to null (meaning "not checked", distinct from false), and the score is computed from the AI Overview alone.
• Plan-limit guard: on FREE accounts the run stops before the first paid request when the forecast exceeds your plan limit — nothing is charged beyond your balance.

What this actor does not do

• It does not rewrite or generate content, keywords, or meta tags.
• It does not crawl or audit your website.
• It does not report classic organic rankings or positions.
• It does not support per-keyword geo or language targeting in v1 — everything is checked against google.com in English.
• It does not return raw search-engine payloads — only the processed, decision-ready fields above.
• It does not measure search-volume trends over time (that is a different job) — it measures AI click cannibalization right now.

FAQ

What is an AI Overview disruption score?

A single 0–100 number per keyword that combines AI Overview presence and format, AI Mode presence, search volume, and how concentrated the AI's citations are. Higher means AI is intercepting more of the clicks for that query.

How is the triage class decided?

Directly from the score: defend at 70 or above, monitor from 40 to 69, and deprioritize at 39 or below. The thresholds are fixed so the classes are comparable across runs.

What is the difference between AI Overview and AI Mode?

The AI Overview is the AI-generated block that appears above Google's organic results. AI Mode is Google's conversational answer experience. This actor checks both and merges them into one row.

Do cache hits cost anything?

No. A keyword scored within the last 24 hours is served from cache with cache_hit: true and is never billed beyond the one-time run start.

How often should I run it?

About every 7 days keeps your triage current without over-spending, since AI Overview presence for a query is relatively stable week to week. You can also run it via the Apify API or the MCP server.

Related actors

Pair this with the rest of the doesaiknow suite for a fuller AI-search and SEO view — local pack and AI Overview tracking, and listing optimization. See the publisher page for companion actors.

Legal & support

• This actor reports publicly visible Google AI Overview and AI Mode results for the keywords you provide. Data is sourced from stable vendor APIs, not from scraping, proxy rotation, or anti-bot circumvention. It stores no personal data.
• Output is provided as-is for competitive-intelligence and SEO-monitoring purposes. Search results and AI answers change constantly; figures are point-in-time observations, not guarantees.
• Google is a trademark of Google LLC. This actor is not affiliated with, endorsed by, or sponsored by Google.
• You are responsible for using the output in line with the terms of the platforms you operate on and applicable law in your jurisdiction.
• Questions or a bug? Open an issue on the Actor's Issues tab, or see the API tab for integration details.

SEO keywords

Google AI Overview, AI Overview monitor, AI Overview SEO, AI Overview tracker, AI Mode SEO, AI search optimization, AI SEO tool, click cannibalization, keyword triage, SGE tracker, generative search optimization, AI disruption score, SERP AI monitoring, keyword portfolio audit, AI Overview citations, search volume triage, AI Overview presence, defend keywords, SEO monitoring, AI answer tracking, Google AI answers, content strategy SEO, organic click loss, AI cited domains.

Built by doesaiknow.