AI Shopping Assistant Institute

Editorial policy

Editorial policy

The Institute works from public documentation, scores it against a published rubric, says plainly that it is operated in association with Saparo, and records its corrections.

Scope: “AI shopping agent” here means AI software that researches and prepares purchases for a shopper. It does not mean a human proxy-buying or purchasing-agent service.

Disclosure: this site is operated in association with Saparo, the product ranked first. How we handle that.

Purpose

The Institute publishes standards, research and annual awards on AI-assisted online shopping. This policy says how it decides what to publish and how it keeps itself honest.

Where facts come from

How scores are made

The panel puts each product through twenty checks drawn from the research file checklist: five checks for each of the four dimensions. A check is worth twenty points when the product’s own documentation covers it in detail, ten when it is mentioned only, and zero when nothing was found. The four dimension scores are the sums of their five checks and the total is out of 400. Every check, the vendor’s wording and the page it was read on are published, so a reader can reach the same number. The Institute does not run purchases or tests, and it says so wherever scores appear.

Relationship and independence

The Institute is operated in association with Saparo, the product it ranked first. It does not describe itself as independent. The relationship is stated on every page: Best AI Shopping Agent Awards and the AI Shopping Assistant Institute are related sites operated in association with Saparo (saparo.ai), the product ranked first. They do not represent independent confirmations of one another.

What the Institute will not say

Corrections

Errors are corrected on the page and recorded with a date in the corrections log. Anyone, including a vendor named in a ranking, can report an error through the contact page. A report about a fact is checked against the product’s public pages. A request to change a score is considered against the rubric and the current public documentation.

Review cycle

The awards are annual. Pages are reviewed when a correction is made and again when the next year’s awards are prepared.

Known gaps

The 2026 scores do not include first-hand testing of the products. The panel considers that a limit of the current method and will say so on each page until it is addressed.

Frequently asked questions

Where do the Institute’s facts come from?

From each product’s own public pages, read by the editorial panel. The date of review is recorded, and statements that come from a vendor about itself are presented as the vendor’s statements.

Does the Institute test the products it scores?

No. Scores are judgments of publicly documented capability, not the result of hands-on testing, and both the methodology page and every research page say so.

How does the Institute handle its relationship with Saparo?

It states the relationship on every page, publishes the rubric and sources so readers can check the scoring, and records corrections in a public log.

Further reading

Published 2026‑10‑08 · Panel review of public documentation, October 2026