Agency AI Visibility Reports: white-label weekly client packets

$12 per client packet. $8 each for books of 20-49 clients, $7 each at 50+. The Store price box shows "$12,000 / 1,000 packets": that is the same $12 per packet. Free quote first; only completed packets are charged.

Paste your client list once. Every week, each client gets a finished, white-label report showing whether ChatGPT, Perplexity, Gemini and Google AI Mode recommend them, which competitors they name instead, which sites they cite, how the client ranks on Google, what changed since last week, and a prioritized action queue with quoted evidence. You also get one portfolio page for the whole book.

It replaces the Monday-morning work between "we ran the checks" and "the client got a report": scoping, collecting answers, matching brands, comparing weeks, writing recommendations and formatting.

• 📦 One packet per client, ready to send: HTML + PDF with your agency name, logo and color. No Wrenfield branding.
• 🤖 4 AI search surfaces: ChatGPT (with web search), Perplexity, Gemini and Google AI Mode. Optional: Google AI Overviews, Claude, ChatGPT API.
• 📋 Where to get listed first: the directory and list pages AI engines cited when they named a competitor instead of your client.
• 📊 Scores that hold up: AI visibility, share of voice, top-pick rate, own-site citation rate, by engine and per competitor, with week-over-week change.
• ✅ Action queue, not just charts: up to 8 ranked actions per client ("Get into the answer for …", "Win the top pick from …", "Get your own page cited …"), each with the quote and the sources the engines cite.
• 🔍 Evidence for every number: a question × engine grid and the quote behind each answer.
• 💵 Price shown before anything runs, charged only for completed packets. Same-week re-runs are free.

Sample packet

How it works

1. Paste your client book as CSV (export it from Google Sheets). One row per client:

Column	Required	Example
client	yes	Help Scout
domain	yes	helpscout.com
brand	no (defaults to client)	Help Scout (add variants with |)
competitors	recommended	Zendesk | zendesk.com; Intercom | intercom.com (separate with ;)
questions	one of these two	up to 20 buyer questions, separated with ;
category	one of these two	help desk software (we write 6 neutral buyer questions)
keywords	no	up to 10 Google keywords to rank-check, ;-separated
location, language	no	United Kingdom, English

2. Add agency name, logo URL and accent color.
3. Click Start. The run first writes a QUOTE (clients, checks, maximum price). Turn on Quote only to see it without running.
4. Open PORTFOLIO.html (Output tab) for the book, and each client's packet from there or from the Client packets table.
5. Save as a task and schedule it weekly. Each run compares every client with the previous week (same Book ID).

A 3-client book with 8 questions × 4 engines and 3 keywords each finishes in about 3 minutes.

Pricing

One charge per completed client packet. The price per packet depends on the size of the book in the run:

Client book in the run	Price per client packet	4.33 weekly runs a month
1-19 clients	$12	5 clients ≈ $259.80 / month
20-49 clients	$8	20 clients ≈ $692.80 / month
50+ clients	$7	50 clients ≈ $1,515.50 / month

Each packet covers up to 20 questions on up to 4 engines and up to 10 Google keywords, plus the HTML + PDF, evidence and action queue.

Not charged: packets where fewer than 90% of checks came back (delivered free, marked partial), re-runs of the same book in the same week (served from the stored packets; turn on Re-collect this week to ask again, which is charged), quote-only runs. Set Maximum cost per run in Apify to cap a run; clients beyond the cap are skipped, not charged.

Free Apify plan: up to 3 free preview packets a week, each for 1 client (5 questions, 2 engines, no rank checks), not charged.

How this compares: subscription GEO trackers charge per prompt per month (for example Otterly lists $189/month for 100 prompts across four engines, checked Oct 22, 2026) and leave the report writing to you. This Actor is priced per finished client deliverable instead.

What's in a packet

Section	What it shows
At a glance	AI visibility, share of voice, top pick, own site cited, Google top 10, each with change vs last week
Action queue	Up to 8 actions, highest impact first, each with why, what to do, quotes and cited sources
Question coverage	Grid of every question × engine: the client's position and who was named first
Brands in AI answers	Client and each competitor: named in, share of voice, top pick, average position, citations, change
By engine	The client's numbers per engine
What changed	Gained and lost mentions, position moves, citations, Google moves
Google rankings	Position in the top 20, AI Overview present, client cited in it, competitor positions
Where to get listed first	Directory and "best of" list pages cited in answers that named a competitor but not the client: the listings to pitch first
Sources AI trusts	The third-party domains cited most across the client's answers (image CDNs and API hosts filtered out)
Evidence	Every answer: named or not, position, first brand and the quote

Output

• Dataset (Client packets view): one row per client with status, charged, visibilityScore, visibilityDelta, shareOfVoice, topPickRate, citationRate, googleTop10, topAction, the full actions list, competitors, whereToGetListed (url, title, domain, answers), and links to the packet HTML, PDF and evidence JSON.
• Key-value store: PORTFOLIO.html, MANIFEST (JSON: per-client status, charge, links), QUOTE, packet-<client>.html, packet-<client>.pdf, evidence-<client> (every answer's brand matches and citations, for your own dashboards).

{
  "client": "Help Scout",
  "status": "complete",
  "charged": true,
  "visibilityScore": 78.1,
  "visibilityDelta": null,
  "shareOfVoice": 22.9,
  "topPickRate": 28.1,
  "citationRate": 34.4,
  "topAction": "Win the top pick from Zoho Desk",
  "reportUrl": "https://api.apify.com/v2/key-value-stores/<id>/records/packet-help-scout.html"
}

API, schedules and agents

Send the book as JSON in clients instead of CSV:

{
  "clients": [{"client": "Acme", "domain": "acme.com", "competitors": "Rival | rival.com; Other | other.com", "category": "project management software", "keywords": "project management software"}],
  "agencyName": "Northwind Digital",
  "bookId": "northwind-weekly"
}

Schedule the saved task weekly in Apify Console, and add a webhook to post PORTFOLIO.html to Slack or email. MCP: https://mcp.apify.com?tools=wrenfield-data/agency-ai-visibility-reports. Call it with quoteOnly: true first to show the cost.

Method and limits

• Questions are asked fresh each run through each engine's live interface or API (ChatGPT with web search, Perplexity Sonar API, Gemini, Google AI Mode) for the client's country and language, signed out, through a licensed data provider. No browser automation, no accounts of yours.
• Brand matching uses the names, variants and domains in your book. Short or everyday-word names (for example Close, Kit, Front) are matched case-sensitively to avoid false hits. Questions that name the client are reported but excluded from the scores.
• AI answers vary between runs. Read a single changed answer as a signal and the weekly trend as the measure. Packets say this to the client too.
• Google positions are organic positions in the top 20 of a desktop search.

FAQ

Is client data shared? No. Runs and stored packets live in your Apify account. We don't read, log or reuse your client list, questions or results.

Can I edit the report? Yes. The HTML is a single self-contained file, and the dataset and evidence JSON hold every number if you'd rather use your own template, Google Sheets or Looker Studio.

What if an engine is down? Each check is retried a bounded number of times. If a packet still has fewer than 90% of its checks, it is delivered marked partial and not charged; re-run it later.

Can I run more than 20 questions for a client? Split the client into two rows with different names (for example "Acme (pricing)" and "Acme (alternatives)").

Related tools

• AI Search Visibility Tracker: raw prompt × engine checks for your own pipelines
• AI Answer Scraper: full answers, citations and structured brand lists from 7 engines
• Keyword Search Volume: Google Ads volumes without a Google Ads account
• Competitor Keywords: the keywords any domain ranks for

Feedback and support

Missing a section your clients ask for, or need another country or engine? Open an issue on the Issues tab with the run ID. We reply there, usually within a day.

Acceptable use

• Independent tool; not affiliated with OpenAI, Perplexity, Google or the brands in the samples.
• Don't use it to manipulate AI answers or search rankings.
• Not available from, or for people or organizations in, Russia, Belarus, Cuba, Iran, Syria, Myanmar, North Korea, Crimea or the Russian-occupied territories of Ukraine.