Our Methodology
Last updated: May 1, 2026
Every ranking on AIWriterPick is produced by a deterministic scoring pipeline — same data in, same ranking out. There are no editorial overrides, no sponsored bumps, and no human guesses. This page explains exactly how it works.
1. Data sources
| Attribute | Source | How collected | Refresh |
|---|---|---|---|
| Monthly price | Tool's own /pricing page | Automated scrape + manual verify | Weekly |
| Yearly price | Tool's own /pricing page | Automated scrape + manual verify | Weekly |
| Free tier | Tool's pricing page + signup flow | Manual testing | Monthly |
| G2 rating | G2.com product page (JSON-LD) | Automated scrape | Daily |
| Review count | G2.com product page (JSON-LD) | Automated scrape | Daily |
| Feature flags | Tool documentation + product testing | Manual review | Monthly |
Every data point carries a fetched_at timestamp in our database. If data is stale or unavailable, we display "—" rather than a guess.
2. Fit scoring model
Each tool is scored 0.0–1.0 for each use case using a weighted rule model. The score reflects how well a tool fits that specific use case — not how good the tool is in general. A tool that ranks #1 for Bloggers may rank #5 for Students, because different criteria matter.
Criteria and weights (examples)
| Use case | Top criteria (by weight) |
|---|---|
| Students | Free tier, plagiarism check, grammar check, price |
| Bloggers | Long-form support, SEO mode, templates, G2 rating |
| Marketers | Templates, browser extension, API access, price |
| E-commerce | Templates, long-form, integrations, free tier |
| Freelancers | Price, API access, long-form, free tier |
| YouTubers | Templates, long-form, browser extension, G2 rating |
The full scoring model is deterministic: given the same database values, it always produces the same ranking. We retune weights when we identify systematic mis-rankings through user feedback.
3. Comparison pages
Head-to-head comparison pages (e.g., "Jasper vs Copy.ai") display verified data side-by-side — pricing, free tier, feature flags, and G2 rating. We don't declare an overall winner because the right tool depends on your use case. Instead, we highlight where each tool leads and link to the relevant use-case pages.
The intro and verdict copy on comparison pages is AI-generated from the verified data — never the other way around. The AI is given only the database values and is instructed not to invent statistics, features, or claims.
4. What we don't do
- ✗ Accept payment for ranking positions
- ✗ Estimate or guess data points — missing data shows as "—"
- ✗ Use AI to invent pricing, ratings, or feature claims
- ✗ Republish copyrighted review text from G2 or other platforms
- ✗ Show affiliate links without disclosure
Found a data error?
If you see a price, feature flag, or rating that looks wrong, we want to fix it. Accurate data is the entire point of this site.