Frequently asked questions
Something not covered here? support@locafact.com · we reply within 24–48 hours (business days).
Support and account
How fast do you answer support email?
Within 24–48 hours on business days. Billing and refund questions are answered first.
I paid but did not receive my sign-in email. What now?
Use the sign-in page recovery box: enter the email you paid with and we send a fresh link to that mailbox. Payments reconcile about every hour, so if you paid less than an hour ago, wait for that pass first.
Do you offer a free plan?
No. Every plan is paid; the entry product is the one-time scan, and its full price is credited toward an audit. There is no free plan; the free check is a 3-question probe on one engine, not a scan.
Can my clients see the FoundNow brand?
No - that is the point. Reports are white-label: your brand name, your colors, your logo, with "prepared by" set to your agency. The PDF carries no third-party branding.
What do you do with my clients' data?
We only sample publicly observable AI answers about a business location. We never ask for account credentials, never post or edit anything on any platform, and never act on a client's behalf. See the privacy policy.
Can my whole team use one account?
Yes. An account belongs to your agency - your own employees use it day to day, and reports are white-label regardless. What the Terms rule out is passing raw platform access to another business. Sign-in links go to your account email, so teams typically use a shared agency inbox.
Why is the company behind FoundNow based in Shanghai?
LocaFact is our trading name; the operating entity is Shanghai Yujun Information Technology Co., Ltd., listed on the Contact page and in our Terms. Payments, receipts and payment disputes are handled by Creem, our merchant of record, in USD - you pay and are refunded through Creem. Reports and client data live on Cloudflare infrastructure, and we never ask for your clients' credentials.
Billing
How do refunds work?
One-time reports have a 7-day window with a verified-defect requirement; every subscription's first payment is refundable in full within 14 days, and you can cancel at any time after. The exact terms live on the refund policy page, and the dashboard has a refund request form.
How does the scan credit work?
The one-time scan is credited in full toward an audit. The code is issued to your account after payment — reveal it on your dashboard and enter it at audit checkout, so you only pay the difference.
Methodology - the short version
This is the 300-word version. The full method - sampling, confidence intervals, model versions and engine coverage - is on the methodology page.
Why re-run the questions monthly instead of daily?
Because AI assistants return a different answer on every single run, a one-day snapshot tells you almost nothing: a good or bad day is noise, and reacting to it means chasing randomness. We re-run the same buyer questions every month and compare distributions over time - that is what turns sampling into a trend you can act on and puts last month's fixes to the test. Audits sample deeper (5+ runs per question, with 95% confidence intervals); monthly reports use a lighter repeat (3 runs) built for trend comparison, never for single-number conclusions. Model versions are recorded with every sample, so an engine-side change shows up as a labeled shift in your trend instead of silently moving your numbers.
Which AI assistants do you sample?
The authoritative coverage for any report is the report itself: it names every engine and model version behind its numbers, and no rate is ever attributed to a surface we did not sample. Coverage expands as we add surfaces. The assistants people actually ask local buying questions of today are ChatGPT, Gemini and Perplexity - that is the set our engine support is built around, because public APIs are what make sampling reproducible at scale: fixed questions, recorded model versions, citable answers. That combination is what lets us put confidence intervals on every number instead of quoting one lucky answer. Assistants without a public API are a different story: no reproducible method, no report.
What about Google's AI Mode and AI Overviews?
Not in v1. There is no public API for them, and we will not scrape unofficially: numbers you sell to clients have to come from a method you can explain to the client. It is on the roadmap - we are tracking the surfaces and will add them when there is a compliant way to measure them. The audit reports already say this explicitly rather than quietly omitting it, so you can set expectations with your clients today.