Chat Moderation

Chat moderation is the safety net around driver messaging. It holds messages with objectionable language before anyone reads them, collects the ones people report, and gives you one place to decide what happens next.

Two different things put a message in front of you.

The first is a word filter. Your company has a list of blocked words, and a message matching one is stopped and held for review instead of being delivered. This is a straightforward text match, not an AI judgment call: it catches the terms on the list and nothing else. If your company has never customized the list, one default profanity term applies, so the filter always has something to catch. The list runs on messages that pass through VoiceERP on their way out: a driver's text-message replies, and anything sent from the manager app as yourself, as Viki, or as a persona. Messages a driver types directly in the app go straight to the messaging platform, where its own always-on filter applies instead. That filter rejects a matching message at the moment the driver tries to send it, so there is nothing to hold for review.

The second is people. Any driver or manager can open a message in the driver app and choose Report message or Block user. Each of those puts an entry in your moderation queue so a human reviews it, even when the person who reported it has already hidden the message for themselves. A manager can also Remove a message or Ban its author directly from the driver app. Those two act immediately and do not create a queue item.

Open Chat, click the ⋯ button ("More chat tools") at the top of the conversation list, and choose Moderation. That menu item only appears when something is actually waiting — it carries a count badge showing how many items still need a decision, and it disappears once the queue is clear.

The page is titled Moderation review and lists each item as a card, newest first, with anything you have already released sorted to the bottom. If nothing is waiting you'll see "No messages are waiting for review."

Every card shows the full message text, who wrote it, the conversation it belongs to, and the time it happened in your company's timezone. Under the author's name you'll see how to reach them — a driver's phone number or a manager's email — so you're never looking at an anonymous account. The card also states why the item surfaced: "Matched profanity filter," "Manually flagged," "Reported by a user," or "User blocked the author."

A red Blocked tag means the message was caught by the filter and never delivered to anyone. An amber Flagged tag means it did go out and was surfaced afterward for a second look.

Active items offer two buttons. Release clears the hold and marks the item Released, which is how you dismiss a false positive — a driver saying something blunt but harmless. Remove takes the message down: a delivered message is taken out of the conversation and a deleted-message placeholder stays in its place, and a blocked message that never went out is simply dismissed as correctly withheld. The item is then marked Removed.

A third button, Ban author, sits on every card and targets the person rather than the message, so it stays available even after you've released or removed the item. It asks you to confirm, then stops that person from posting anywhere in chat. The ban is reversible, but lifting it isn't something you do from this page — ask your VoiceERP contact to unban them.

Each item can only be handled once. If a colleague releases or removes something a moment before you do, your click is rejected rather than quietly overwriting their decision, so refresh the page to see where things landed.

A driver whose text message is caught by your word filter doesn't get a warning or a bounce-back. The message never appears in the conversation, and nobody else sees it. A driver who types a matching message in the app sees the send fail right away, and the message is not delivered.

A driver who reports a message is told "Thanks for letting us know. Our team will review this message." A driver who blocks someone stops seeing that person's messages immediately and is told your team has been notified; the blocked person is moved into a Blocked section they can unblock later. If a manager removes a message, everyone sees a deleted-message placeholder where it was.

Moderation is on by default for every company. Two things are configurable: whether moderation runs at all — which controls the word filter, the Report and Block options, and the manager-only Remove and Ban actions in the driver app — and the list of blocked words itself. Neither is edited from the manager app; ask your VoiceERP contact to turn moderation off or to adjust your terms. Turning it off entirely only makes sense for companies using chat purely for announcements, where drivers never post.

Moderation is one of several chat tools in that same ⋯ menu, alongside templates, scheduled messages, archived chats, and custom emoji — each covered in its own article.