Where verification lives
Verification lives inside your existing Submissions page. It is not a separate place to visit. The engine does the routine work, and the queue is organized so your attention goes only to the submissions that genuinely need a person.
The Submissions sidebar entry carries a needs-attention badge: a count of everything still open (waiting for the engine, needing your review, or verified but not yet rewarded), so you can see unfinished work without opening the page.
The tabs
Submissions are grouped into tabs by where they are in their lifecycle:
- Need action: everything still open, whether it is waiting on the engine (Pending), needs your review (Manual review), or is verified but not yet rewarded. This is your working queue.
- All: every submission, regardless of stage.
- Verified: verified, by the engine or by you, with the reward not yet fulfilled.
- Rejected: rejected, automatically or by you.
- Rewarded: verified and the reward has been delivered. This is the final stage; there is no reversal.
Verified, Rejected, and Rewarded do not care who decided. A submission you approve by hand lands in Verified exactly like an auto-pass. Who decided is recorded in the submission’s history, not in the tab.
The list
The table for the active tab has these columns: Person (name and email), Campaign, Platform, Source, Status, AI verification score, Reward, and Submitted.
- Status shows where the submission sits: Pending (still in the queue), Manual review (waiting for you), Verified, or Rejected.
- AI verification score shows the 0 to 100 number, or N/A when a submission was not scored (for example an older manual submission).
- Reward shows the reward attached to the submission.
You act on a submission by clicking its row, which opens the detail drawer. Decisions are made there, not inline in the table.
Bulk selection. Each row has a checkbox, plus a select-all checkbox in the header, so you can act on many submissions at once from the toolbar that appears when rows are selected.
You can also search by email and filter by campaign, platform, and source from the controls above the tabs.
The detail drawer
Open any row to review it in a two-column drawer. The layout is the same for every status, so you always know where to look.
Left, the submission:
- The advocate: avatar, name (a link to their person record), email, and their company when known.
- A View on link, named for the platform (for example View on G2), that opens the live review or post on its platform, so you can check the source in one click.
- The proof itself, under a Screenshot proof heading (or a video player or written quote, depending on the proof type).
Right, the verdict and what to do:
- Decision buttons: Approve & Reward and Reject sit at the top.
- The verdict: a score ring (for example 20 / 100), the status (for example Manual review), a one-line reason (for example “no author found”), and an “X of Y checks passed” summary.
- How the score was reached: the list of checks the engine ran, each with a pass or fail marker. Hover any check for details. A How it works link points to How AI verification works.
- Re-run AI verification: re-runs the engine on this submission.
- Details: campaign, platform, section, source, reward, and submitted date.
- History: every event, stamped with who and when. A human action shows the acting person’s email; an engine event shows AI. So you can always tell an automated step from a human one.
In manual mode, or on a submission that was never scored, there is no verdict block. You review the proof on the left and decide on the right.
The decisions you can make
From the drawer you can:
- Approve & Reward: approve the submission and release its reward. Under the Hold release policy the verified reward then waits in Rewards → Activity until it is released; under Auto it goes out on its own. See How rewards work with verification.
- Reject: reject the submission. You can add a reason, which is shared with the advocate. A reason is optional, but the action is always recorded against you.
- Re-run AI verification: re-run the engine on this submission, for example after a not-yet-live review has had time to publish.
Every decision records who acted (your email) and, on a reject, the reason. The engine’s reason is always shown alongside an AI verdict. This is what appears in History.
Clearing an author mismatch
When a submission comes to you with the reason no author found or a name that does not match, the engine could not tie the name on the review to this advocate. Advocates legitimately use different name forms across platforms (a nickname, initials, a maiden or married name), so the engine routes these to you rather than rejecting outright.
Open the submission, use the View on link to check the source, and if you recognize the person, Approve & Reward. The engine remembers the name forms tied to a real advocate over time, so repeat submissions from a known name need less re-checking.
If a name matches a different advocate, the engine rejects it and flags it, rather than routing it to you.
What every status means
The Status column shows a small set of values. The specific cause is in the drawer’s one-line reason.
| Status | Means | What to do |
|---|---|---|
| Pending | Waiting in the queue: still in the cooldown before verification starts, still being checked, or a valid review that is not public yet on a platform that moderates posts (re-checked over the following days). | Nothing yet. The engine is still working. |
| Manual review | The engine finished with an unclear result and handed off, or the campaign is manual. | Open it, read the reason, check the source, and decide. |
| Verified | Approved by the engine or by you, reward not yet fulfilled. | Release or fulfill the reward if needed (see How rewards work with verification). |
| Rejected | Rejected by the engine or by you (including a link already claimed by a different advocate, or content that was removed or invalid). | Nothing, unless you want to override it back to verified. |
Common reasons you will see behind a Manual review status include: no author found or a name that does not match, a social post with no product or company mention, a link that disagrees with the screenshot, or a page the engine could not read (blocked, private, or login-walled).
A high score does not guarantee an auto-verify: those safety checks can still route a strong-scoring submission to you. The reason is always on the submission. For the scoring detail behind these, see How AI verification works.