Methodology
Listed vs Tested
Listed: the name, description, URL and pricing come from the tool's official website, with the date we checked. Listed tools get no score, and their pages are not indexed by search engines until a test is done.
Tested: our agent signed up (free tier or trial where one exists), ran a standard task, and recorded what happened.
How hands-on testing works
- The agent creates an account, using the free plan or trial where one is available.
- It runs a standard task for the tool's category (for example "build a lead-enrichment workflow" or "fix a failing test in a sample repo").
- It records timings, the steps needed, the real cost, failures, and screenshots.
- It scores against the rubric below and publishes the evidence, including failures.
- It re-tests when a tool changes significantly, and every review shows its test date.
Scoring rubric (out of 100)
| Criterion | Weight | What we look at |
|---|---|---|
| Task success | 30 | Did it complete the standard task correctly? |
| Autonomy & reliability | 20 | How much hand-holding did it need, and how often did it fail or stall? |
| Setup & ease of use | 15 | Time from signup to first useful result. |
| Value for money | 15 | Real cost of the test task, and what the plans include (GBP/VAT noted where relevant). |
| Safety & control | 10 | Approvals before consequential actions, permissions, data handling (UK GDPR notes). |
| Integrations | 10 | Connects to the tools a UK SMB actually uses. |
Honesty policy
- Sponsors never affect rankings or scores. Paid listings, sponsorships and affiliate commissions buy visibility, which is always labelled Ad, never a better score.
- We never invent scores or test results. No test, no score.
- A paid "tested review" pays for the test to happen sooner, not for the result. Bad results get published too.
- Corrections are welcome at hello@agentscanr.com.