Inbox Sidebar Actions

Manage visitor interactions and conversation controls through the sidebar actions panel.

The Details panel

Beside an open conversation, the Details panel shows who the visitor is and what state the conversation is in. Where the screen is wide, the panel button at the top right of the conversation hides or shows it; on a narrow screen the same button swaps the conversation for the details.

From top to bottom it holds:

  • The visitor card — the visitor's name, email, a green or gray dot for online or not, and an Edit button (see Editing Visitor Information)
  • The conversation block — Status (Open, Waiting for reply, Answered, Resolved, or Snoozed until …), Assignee, Brand, Channel, Started and Last activity
  • The Profile rows — the visitor's Id, Memory with a Show button (see Visitor Memory), Plan, Country, External ID, Last seen and, where your workspace has the Security feature, the Trust score. Rows with nothing in them are left out
  • A block card — only while the visitor is blocked (see Ban visitor below)
  • Folding sections — Customer data (whatever your website sent along when it identified the visitor; it appears only when there is some), Recent activity (see Recent Events), Device details (widget, IP, location and device) and Actions, which starts folded

Available Actions

The actions below are in the Actions section, in this order: Related email chats, Ask for rating, Send transcript, then — after a thin line — Ban visitor (or Unban visitor) and, for owners only, Delete visitor data.

Shown when the visitor has an email address. It opens Related Email Chats ("Other chats in the same widget with this email address") with the other conversations of the same address, and a click on one opens it. When there are none, the dialog says "No other chats found with this email address".

Ask for rating

Request feedback from the visitor about their conversation experience.

How it works:

  • Sends a rating request that appears to the visitor in the chat widget
  • Works on Facebook Messenger and Instagram conversations too. There the visitor gets five tappable star buttons (⭐ to ⭐⭐⭐⭐⭐) instead of the in-widget rating card, and their tap is recorded the same way. The question the visitor sees is also stored in the conversation, so the thread reads question → stars
  • Once the visitor has rated you in a chat, you can't ask again in that chat for 7 days
  • The request is recorded as a private event message in the chat
  • The button is only available on the conversation you are assigned to, and is not offered on phone calls. When it is grayed out, hovering it says "You cannot ask for rating on this conversation" or "The visitor left a review in the past 7 days"

When to use:

  • After resolving a visitor's issue
  • When you want to track conversation quality
  • To gather feedback for service improvement
The rating request is disabled if the visitor has rated this agent in this specific conversation in the past 7 days. With the widget's Ask for a rating switch on (Chat widget → Chat tab, After a chat), the visitor is also asked automatically when the agent leaves the conversation, if that agent sent at least 2 replies and the visitor isn't waiting for an answer.

Send transcript

Email a complete transcript of the conversation to the visitor.

How it works:

  • Retrieves all public messages from the conversation
  • Groups messages by date for better readability
  • Sends a formatted email containing the full chat history
  • Adds a public event message confirming the transcript was sent
  • Goes only to the visitor's own address — the one the visitor gave: in the widget's email form, in answer to a workflow's question, in the roadmap's email box, through your website when they are logged in there, or as the sender of an email they wrote to you. The dialog shows it as fixed text; there is no recipient field
  • When the visitor has given no address, Send transcript is grayed out, and hovering it shows "This visitor has not given an email address."
  • A transcript is an email, so it follows the email rule: free with a paid plan or during your free 14 days; otherwise it costs 1 credit from the credit you have bought. With none, it is not sent, no "transcript sent" line is written, and the payment window opens
Only the visitor can give or change their email address. In the visitor's details (Editing Visitor Information) the email is shown as read-only text with "Only the visitor can give or change their email address." The name and the other fields stay editable.

Configuration options (in the Send transcript dialog, which also shows the fixed To email line):

  • From email: Select from verified organization email addresses
  • From name: The name that appears as the email sender. It is hidden while From email is the shared address [email protected], which sends under your brand's name (or your workspace's name if there is no brand name)

Use cases:

  • Providing visitors with conversation records
  • Answering a visitor who asks for the conversation by email — visitors can't send themselves a transcript from the widget

To email the visitor a single reply instead of the whole conversation, switch on Also send by email next to the send button in the message box — see Inbox.

Limits:

Transcripts have limits per conversation, per organization and per email address, so they can't be misused. Over a limit the transcript is not sent, and the error message says which limit was reached.

Ban visitor / Unban visitor

Block a visitor from your team — at one of two levels — and manage that block right from the conversation.

Two block levels

When you ban a visitor you choose how far the block goes:

LevelWhat the visitor can doWhen to use
Block from agents only (1)Keeps the AI assistant but can't reach a human. The AI won't connect them to an agent — and won't pretend it did.Someone who shouldn't take up agent time but can still self-serve
Block completely (2)Can't use the chat at all.Spam, abuse, or anyone who shouldn't reach you at all

How it works:

  • Ban visitor opens a short dialog ("Block this visitor from messaging. Add a reason so your team knows why.") where you pick the Block level, enter a Reason (required), and choose a Duration — 1 hour, 24 hours, 3 days, 7 days, 30 days, or Forever
  • Timed bans lift on their own — when the duration ends, the block simply stops applying; nobody has to remember to unban
  • The visitor is never told they're banned and never sees the duration — when a timed ban ends, chat just works again for them
  • If the visitor was banned before, the dialog lists their previous bans — level, duration, reason, date, and whether each one was lifted or expired — so repeat offenders are obvious at a glance
  • The ban is logged with the reason, who applied it, and when, plus a private event note (agents only, never the visitor)
  • The change takes effect immediately across the visitor's active sessions
  • Ban and unban events applied from a conversation are recorded against your organization, so they also show up in Recent activity (as Visitor banned and Visitor unbanned, the ban with its reason) and in the activity timeline of the Visitors page (Audience → Visitors, Visitor details & activity)
  • A visitor in your own organization can be banned even if they have never actually opened the chat widget and have no conversation yet — with no conversation there is nothing to record the event against, so that ban leaves no entry in the Visitors page activity timeline
  • Block from agents only needs an AI to keep the visitor company. On a conversation where Yaplet does not answer — the brand has no AI agent, Enable Yaplet is switched off, or its Answers on switch for that channel (Brand → AI agent → General) is off — a level-1 ban is applied as a full block instead
  • You can also ban or unban a visitor from the Yaplet mobile app, with the same level and duration options

Also block their connection (full block only)

A full ban can also block the visitor's connection — the internet address their browser connects from (for IPv6, the household-sized block of addresses) — so they can't simply open a private window, or clear their site data, and come back as a "new visitor":

  • The Also block their connection box is ticked by default on a full ban. Untick it to ban only the visitor
  • The connection stays blocked for as long as the ban, capped at 24 hours — a 1-hour ban blocks the connection for 1 hour; anything from 24 hours up to Forever blocks it for 24 hours. Internet addresses get reassigned, so a block that outlived a day could land on a stranger
  • It only stops new visitor sessions and new conversations from that connection. Anyone already mid-conversation on the same connection — a colleague on the office Wi-Fi, another customer behind the same mobile network — is not affected
  • The box is only offered when the visitor was actually seen on that connection within the last 24 hours. If their last visit is older, the box is disabled with a hint: the connection may belong to someone else by now, so only the visitor is banned
  • A visitor turned away by a connection block sees the same quietly disabled chat input as any banned visitor — nothing more
  • Lifting the ban releases the connection block too. Every blocked connection is also listed on the Network tab of the Security page, where the connection part alone can be released
  • An agents-only (Level 1) ban never blocks a connection, and the Yaplet mobile app currently bans without the connection-block option

Ban the others on that connection too

With the box ticked, the dialog also lists the other visitors seen on the same connection in the last 24 hours — name or email (or Anonymous visitor), country, when they were last seen, and their last message — all pre-selected. Confirming applies the same ban (level, reason, duration) to every selected one in the same click, so a spammer with five browser windows open is dealt with once. Use Select all / None to adjust, and untick anyone who is obviously a different person. The list shows up to 200 visitors and says when more are included. The confirmation message spells out what happened — visitor banned, connection blocked, and how many more were banned.

Managing an existing block:

While a visitor is blocked, the Details panel shows a block card with a Full block or Agents blocked · AI on badge, the reason, Banned by, the expiry of a timed ban (Expires), and quick controls — Escalate to full block for a Level 1 block, Downgrade to agents-only for a full block, and Unban visitor to lift it entirely. Changing the level keeps the remaining time — escalating or downgrading never extends or shortens the clock. In place of the reply box, the conversation shows Replying is paused — this visitor is blocked with a Lift block button. A block set by the Security system instead says so, and offers Review in Security (for those who may open it) — clear that one there. You can do the same per row on the Banned visitors tab under Audience → Visitors (Visitors), where each row has Unban and, under ⋯, Block completely or Block from agents only.

A blocked chat locks agents out. While a visitor is blocked (either level) you can't send messages, assign the chat, or start a workflow on it — lift the block first. Applying a Level 1 block mid-conversation hands the chat to the AI: the agent is released and the assistant takes over with a short, context-aware message telling the visitor a human isn't available but it can still help. If the widget has no AI assistant, a Level 1 block applies as a full block instead — there's nothing for the visitor to fall back to.

Effects of a full block (Level 2):

  • The visitor can't start or continue conversations — their chat input is simply disabled, with no mention of a ban or its duration
  • Existing conversations stay readable for the visitor and accessible to your team
  • Proactive Engagement chat messages aren't sent to them (banners, surveys, and tours are unaffected)
  • The visitor is listed on the Banned visitors tab under Audience → Visitors (Visitors)
  • If this is a caller's phone profile (the visitor of a phone-call thread), their calls are rejected before pickup — no ring-through, no greeting, no per-minute billing
The visitor is never told they're banned — chat quietly appears unavailable instead. This is deliberate: announcing a ban invites retaliation and immediate evasion with a new browser session, while a quiet block defuses the situation. For heat-of-the-moment abuse, a timed ban is usually the better tool than a permanent one.
Voice calls. Every caller has their own phone profile: their first call creates it, and every later call from that number lands on it. When that phone profile is fully blocked, the caller is hung up before the call connects. Block the caller from their call thread: banning a website visitor whose profile merely shows the same phone number (for example one your website passed in with identify()) does not block calls. Timed bans apply here too — once the ban expires, the next call connects normally. The block is per organization — a visitor blocked in your workspace can still reach other workspaces' numbers. First-time callers without a visitor record yet can't be pre-blocked; the block kicks in from their second interaction onward. A Level 1 block doesn't affect voice — voice has no live-agent handoff anyway, so the caller still reaches the AI.

Delete visitor data

Only the workspace's owner sees this action. It asks twice — first "Are you sure you want to erase this visitor? This will permanently delete the visitor and all chats.", then "This action is irreversible. Confirm that you want to permanently erase all data for this visitor." — and then removes the visitor, their conversations and their memory for good.

© 2026 Yaplet. All rights reserved.