Article
LinkedIn automation vendors make stronger claims than any other corner of the GTM stack, and their buyers carry the most risk. The account on the line belongs to you or your client, not the vendor. That asymmetry shapes our whole protocol for the category.
Before a score exists, we capture everything the vendor publishes. The marketing site, every help center article, the API reference, the terms of service, the privacy policy and the subprocessor list. We also unpack anything they ship publicly.
The extension audit is the part nobody else does. A Chrome extension declares its permissions in a manifest file anyone can read. When a tool asks for access to every website you visit rather than just LinkedIn, that fact goes in the review, with the manifest archived as evidence. It takes half an hour and settles arguments that marketing pages never will.
Then come the contradiction checks. Fifteen document pairs, run on every review. The two that produce findings most often: a safety claim on the homepage next to a restriction-recovery guide in the help center, and a “full API” claim next to an endpoint list you can count on one hand.
The scores follow the documents. Eight dimensions, fixed anchors, and a confidence flag on each. Where the record is silent we say “not documented” instead of guessing, and every vendor sees the findings before publication.
If you run one of these tools and think we’ve read your docs wrong, the corrections log is where that conversation happens in the open.