Questions people ask before they run a scan

Language models sample their output, so the same question can produce a different list of businesses on different runs. Anserra takes several samples per prompt and reports a range rather than a single number. A change is reported only when it moves outside the previous range.

The free scan takes one sample per prompt across five prompts. Paid plans take 3 samples per prompt across 25 prompts, weekly. The methodology page explains how the confidence interval is calculated.

No. Answers are collected through the providers' official APIs. That keeps data stable and within each provider's terms. It also means the answers are those of the model, not of a particular logged-in user's chat history.

Yes. Each business starts with a generated prompt set for its niche and city, and you can edit it within your plan's prompt limit.

The score combines mention rate, accuracy of what is said, source coverage and agent readiness into one number from 0 to 100, always shown with its interval. The report shows each component separately, so you are never left with only the number.

The free scan finishes in about 60 seconds: it samples up to 12 pages of the site and respects robots.txt. A full paid scan with an agent run takes longer, depending on the site and the queue; you get an email when it is done.

A real browser agent opens your site and tries five tasks a customer would delegate to an assistant: understand the offer, find the price, find the booking path, check the conditions, reach a human. It reports the exact step where it stopped, with a screenshot. No other AI visibility tool lists this.

No. The agent follows links that already exist on the page. It never types into fields, never clicks submit, never creates a booking, order or payment. Reaching the booking or checkout address counts as success. This is written into our Terms of Service, section 3.

Add a Disallow rule for the AnserraBot user-agent in your robots.txt. The AnserraBot page explains the user-agent string and rate limits.

Levels from the open Agent-Ready Local Services specification. L0: not readable by agents. L1: readable — offer, contact and hours as text. L2: bookable — an agent can reach the booking step. L3: interoperable — machine-readable availability via /.well-known/agent-service.json. The L0–L2 validator is free.

One business: Solo, $79 a month. Up to three locations or businesses with competitor tracking: Plus, $199. Agencies: Multi-15 ($799, 15 businesses) or Multi-50 ($2,290, 50 businesses, weekly agent runs, API). More than 50: Enterprise.

The free scan is the trial: one business, no account. Paid plans are billed monthly from the day you subscribe. If you cancel within 14 days of payment and have not run a paid scan, you get a full refund under the Refund Policy.

From your account, any time. Access continues until the end of the paid period; there is no cancellation fee. Details in the Refund Policy, section 7.

Payments are processed by Paddle.com Market Ltd as Merchant of Record, so refunds are issued by Paddle. Grounds: cancellation within 14 days of payment before any paid scan; a technical failure that fully prevented use during the paid period, reported within 14 days; duplicate or erroneous charges; and cases required by law where you are. Requests are reviewed within 10 business days. Full text in the Refund Policy.

Every plan has a hard cap on the number of AI-engine responses collected per month, derived from prompts × engines × samples × weekly runs × businesses. It is shown in your dashboard. When you reach it, runs pause until the next cycle; you are never billed for overages without asking.

The base plans cover ChatGPT, Perplexity and Google AI Overviews. Claude and Gemini can be added per business for customers who want those engines tracked too: $19 per engine per business per month, same prompts and samples as the base engines.

Enterprise is for more than 50 businesses. You pay the Multi-50 price ($2,290 for the first 50) plus $39 per additional business per month for businesses 51–100, $34 each for 101–250, and $29 each from 251. Example: 100 businesses = $4,240 a month. The full sheet is a PDF on the pricing page; invoice billing is available.

Prices are in US dollars. Paddle calculates and collects applicable VAT, GST or sales tax at checkout based on your location, and issues the invoice.

Not yet. All plans are monthly. Enterprise can be invoiced on custom terms.

Yes, up or down, from your account. Moving up takes effect immediately with a prorated charge handled by Paddle; moving down takes effect at the next renewal.

Open the bulk scan page, paste domains one per line or upload a CSV with your clients' domains (business name, city and vertical are optional). Preview the rows, confirm, and the batch goes into the queue. Results appear per domain as they finish; you get an email when the batch is done. Save the list to re-run weekly. Limits: 5 domains per upload on Plus, 15 on Multi-15, 50 on Multi-50.

PDF reports, client result links, bulk-scan exports and forwarded notifications carry your logo and colors. The report's substance stays as generated. Your clients do not receive Anserra accounts or API keys through your plan.

Yes. Agency plans include multi-user access within the plan's seat count.

Yes. Scan a prospect list, show the prospect that AI names a competitor instead of them, and attach the fix plan as your proposal. Prospects count toward your business limit while tracked.

Public data only: your website as an AI crawler sees it, AI-engine answers to public questions, and public listings. Account data is your name, email and company. Payment data is handled by Paddle. Full detail in the Privacy Policy.

No. Anserra reads public websites and public AI answers. It does not access, process or store patient records, customer lists or any special categories of personal data.

Database and authentication in the EU (Ireland); background workers in the EU (Amsterdam); hosting on Vercel in the US; AI-engine queries go to the respective US providers. The Privacy Policy lists every sub-processor and region.

The current snapshot for the duration of the subscription; change history for 12 months; raw AI-engine text for 12 months; 30 days after account closure, then deletion. Free-scan results are kept for 30 days, with an anonymized version retained for research.