AIFinderOne · Transparency
Rating methodology
This page explains exactly what a user rating means, what it does not mean and how AIFinderOne keeps ratings separate from search recommendations, editorial facts and sponsored placements.
01 · Definitions
Two different signals
A user rating describes the experience reported by AIFinderOne users. Search relevance describes how closely a tool appears to match a particular request. They answer different questions and are displayed separately.
- User rating: an average from published 1-to-5 reviews.
- Recommendation order: task-specific relevance for the current search.
- Neither signal is presented as a universal guarantee of quality or suitability.
02 · Calculation
How a user rating is calculated
The displayed rating is the arithmetic mean of published AIFinderOne reviews for that tool. The review count is the number of published reviews included in that average.
- Only ratings from 1 to 5 are accepted.
- A registered account can have one active review per tool; editing it replaces the previous version instead of adding a duplicate.
- Removed, rejected or unpublished reviews are not included.
03 · No data
What happens when there are no reviews
AIFinderOne does not fill empty profiles with provider scores, marketplace totals or estimated numbers.
- The interface shows “No reviews yet”.
- No star average or unsupported review total is displayed.
- AggregateRating is not added to JSON-LD until at least one real published AIFinderOne review exists.
04 · Integrity
Review integrity and moderation
Reviews should describe a genuine experience and remain useful even when they are critical.
- Spam, duplicate accounts, coordinated manipulation and undisclosed incentives are prohibited.
- Reports may trigger moderation; legitimate criticism is not removed merely because it is negative.
- Personal data, unlawful content and content unrelated to the product may be limited or removed.
05 · Search
How recommendation order works
The search system interprets the request and ranks relevant catalog candidates. It is not calculated from the user star average.
- Signals can include direct name match, task/category match, query terms and curated task relevance.
- Requested constraints such as free access, API availability and budget are used as filters when present.
- Insufficiently verified external results are rejected rather than given a fabricated score.
06 · Independence
Pricing, partners and editorial facts
A price is a sourced product fact, not a vote. An affiliate relationship is a commercial disclosure, not a quality rating.
- Payments and partnerships do not raise a user rating or organic recommendation position.
- Sponsored placements must be labelled and visually separated from organic recommendations.
- Provider claims are checked against official sources and do not become user opinions.
07 · Accountability
Corrections and appeals
Users and product owners can report an incorrect price, feature, source, review or classification through the contact form.
- Include the exact profile, disputed field and supporting evidence.
- Confirmed changes are reflected in the profile and verification date.
- Review moderation disputes are assessed under the public Review Guidelines.