Most “anonymous” surveys are anonymous the way an unlocked diary is private — a promise, sitting on top of a database that still has your name next to your words. Forum 360 is built the opposite way. Read how it works, then decide.
There is nothing raw left to subpoena.
Five steps. The third is the one that matters: after it, your original words don't exist anywhere.
During the open cycle, your words are yours. Edit them, rewrite them, change your mind — nobody sees anything yet.
Every comment is rewritten. Your phrasing, your tics, the expression you always reach for — gone.
The moment the report is generated, the raw text is destroyed. Only the reworded version survives.
Your reworded feedback is combined with everyone else's before a single line is written.
Built from paraphrases only. No sentence you typed exists anywhere in it — because it no longer exists at all.
While a cycle is open, your original answers exist so you can edit them. The instant the annual report is generated, every raw answer is run through an AI paraphrase and the original text is deleted. What's left is the reworded version and nothing else. There is no archive of raw answers, no backup table of “what people really said,” no export button that reaches them — because after the report is built, they are gone.
The people in your forum have read your emails for a decade. Writing style is a fingerprint. So before anyone sees anything, an AI rewrites every comment from scratch. Reports are assembled from these paraphrases only, never from anything you typed.
The rewording and report-writing run through Anthropic's Claude API under terms where inputs are not used to train models. Forum 360 doesn't sell your data, doesn't share it, and doesn't train anything on it. Your feedback does one job: get reworded and combined. Then the raw version is gone.
This is the part to be forensic about. The part of the system that writes and displays reports has no path to raw answers. The raw columns are withheld from every browser-facing role at the database layer, using column-level grants and row-level security, not hidden behind a checkbox in the interface. The software that could leak your words is physically unable to read them. Even the person running the cycle can only ever see who has finished — never anyone's words, including their own.
If fewer than 3 members review someone, that report is never generated. And every line in a report is built from at least 3 reviewers' merged answers, so no single sentence points at one person — even in a room of six or eight where everyone knows everyone. A report thin enough to unmask its source simply doesn't get made.
If one member of your forum has voiced a particular view out loud for years, no software can make feedback shaped by it unrecognizable, and any tool that claims otherwise is lying to you. Here's what is guaranteed instead: nothing in any report points at a person — not the wording, not the ordering, not the timing, not a shred of attribution. The mechanics are exact even where human memory isn't.
For the record: data is encrypted in transit and at rest, never sold, and never shared. Raw answers exist for one purpose, to be reworded and combined, and are deleted once the report is generated. Everything else you see above is enforced by the database, not by our good intentions.