Home Conversations | Reply, organize and find tickets

Conversations | Reply, organize and find tickets

How to reply, filter, merge and follow the day-to-day conversations.
Ian Kraskoff
Alexandre Machado
By Ian Kraskoff and 6 others
• 20 articles

View Previous Conversation with a Contact Without Leaving the Support

https://www.loom.com/share/39b1d25270be4db58510cda68b590286 When to Use - When you open a conversation and want to know if that contact has spoken with us before. - When the client seems to bring up a topic again and you need the context without leaving the screen. - When you want to know through which channel the previous contact occurred and how long ago it was. - When the client has returned multiple times and you want to go through their history step by step. How it Works The notification at the top of the conversation When the contact has had a previous conversation, a notification appears at the top of the message list. It is not fixed: you need to scroll to the beginning of the current conversation to see it. The notification summarizes the main details of the previous conversation: - Channel through which it occurred (Webchat, Email, WhatsApp); - Status of the conversation (open, pending, deferred, resolved); - How long ago the client echoed back to us. :::info The status describes how the previous support was concluded, not whether the client's issue was resolved. A conversation marked as resolved means someone closed it — whether the client was satisfied remains your judgment. ::: Opening the previous conversation Click on View history and the previous conversation opens right there, just above the current one. It appears in a read-only mode: you can read the messages, with their date and time, just as in the main conversation, but you cannot reply there. To resume the previous chat, use the "Open Original" link, which takes you to it. To close, click on Collapse. Going deeper into the history If the client has returned several times, each opened history offers the previous one, up to five levels. Each new conversation appears above the previous one, following the timeline: the higher on the screen, the older. After opening a history, the next notification takes a few seconds to appear. This is intentional: it prevents an accidental click from jumping two levels at once. Long conversations When opening a history, you see the 20 most recent messages of that conversation — usually enough to understand how it ended. If you need more context, click on Load more messages. The button disappears when there is nothing left to load. Notes - The previous conversation is always the one before the one you are reading. If you open a conversation from the middle of the history, the notification shows the one before it, not the most recent of the contact. - The notification does not appear in inboxes configured for one conversation per contact. In these channels, the contact always maintains the same conversation, so there is no separate history to consult. - Merged conversations do not appear as previous. Their content is already in the conversation that received them. - If the contact only had this conversation, no notification appears. - The history is only loaded when you click. While the notification is there idle, it consumes nothing.

Last updated on Sep 02, 2026

How to Merge Conversations from the Same Contact

https://www.loom.com/share/5602c948fe844cd6bba093227dc5e96f When to Use - The same client has opened multiple conversations, causing scattered support. - You want to combine duplicate conversations from a contact into one, without losing history. - Conversations are in different channels (WhatsApp, web chat, email), and you want to centralize them. Prerequisites :::info Conversations must be from the same contact, and the contact needs to have at least two active conversations (open or pending). Resolved or already merged conversations are not eligible. ::: - The conversation you are viewing must be open — you cannot merge from a closed conversation. - When the contact has an open conversation in parallel, a warning appears at the top along with the Merge Conversations button. What the Merge Does - You select which conversation continues (the "survivor") and which are closed. - Closed conversations are marked as resolved and linked to the survivor. - The survivor displays a block labeled "Conversation #N merged here" with a View History button that opens, in read-only mode, all messages from the merged conversation. - The closed conversation contains a note with a clickable link to the survivor, and the reply field is blocked. - Optionally, a transition message is sent to the client through the closed conversation channel. Step-by-step Step 1 — Select Conversations In the open conversation, click on Merge Conversations at the top. In the first step, select which of the contact's other active conversations to include in the merge (you can search by number or ID), then click Next. Step 1 of merging: select contact's conversations Step 2 — Choose the Survivor and Transition Message In the second step, select which conversation continues (marked as "Continues"; others show "Close"). Enable the transition message — and edit the text if desired — or disable it. Then click Merge. Step 2 of merging: choose survivor and transition message Step 3 — Review Results The survivor shows a block labeled "Conversation #N merged here" in the timeline. Click View History to open, in read-only mode, all messages from the closed conversation — including the transition message sent to the client. Read-only merged conversation history :::success All data is preserved: the history of the closed conversation remains accessible within the survivor, and the closed conversation links to it. ::: Setting the Account Default An administrator can configure the default transition message for the entire account in Account Settings → Advanced Settings → Conversation Merge Transition Message: - Enabled / Disabled by default — if the message is enabled when the agent opens the merge. - Use default message — keep enabled to use the default text, or disable it to write a custom message for the account. The {channel} token is automatically replaced by the channel of the continued conversation. Agents can still toggle and edit the message during the merge (Step 2) — account settings are just a starting point. Default transition message configuration in Account Settings What if the Client Replies Later? If the client sends a new message in a merged conversation (reopening it), the link is automatically broken, and the conversation returns to normal. Notes - Merge is by contact — it does not merge different contacts. - Works across channels: you can merge a WhatsApp conversation with a web chat of the same contact. - The transition message always points to the channel of the continued conversation. When both conversations are in the same channel, it indicates to continue in that conversation (instead of repeating the channel name); if they are the same conversation in the same channel, no message is sent. - The merged, closed conversation exits from support metrics (and is filterable).

Last updated on Sep 02, 2026

Conversation Filters

Overview The advanced conversation filters in Cloud Chat have become more convenient. Now you can see how many conversations match the applied filters, view which filters are active without reopening the modal, remove individual filters, and sort results even with filters active. Filtered Conversations Count When filters are applied, a counter next to the "Conversations" title shows how many conversations match your search. This helps gauge the volume before starting to work on the queue. Active Filter Chips After applying filters, each filter appears as a chip just below the conversation list header. Each chip displays the attribute and the filtered value (e.g., Inbox: Email or Status: Open). To remove a specific filter, simply click the X on the corresponding chip. The list updates automatically. If you remove the last chip, the view returns to the default unfiltered state. Sorting Available With Active Filters Previously, the sorting dropdown (most recent, oldest, etc.) would disappear when advanced filters were active. Now, it remains always visible, and the chosen sorting is correctly applied to the filtered results. Frequently Asked Questions Does the count include conversations from all inboxes? The count exactly reflects the result of the applied filters. If you filtered by a specific inbox, the count shows only conversations from that inbox. Can I remove all filters at once? You can remove chips one by one or remove all filters at once. Removing the last chip automatically resets the list to the default state. Does sorting work with saved folders too? Yes. The sorting dropdown now remains visible both when advanced filters and saved folders are active.

Last updated on Sep 02, 2026

Copilot: AI Assistant to Improve and Edit Responses (Including Translation)

When to Use - When the agent needs to rephrase, shorten, or expand a draft before sending. - When the agent needs to translate the response into the customer's language. - When the agent wants to adjust the tone (more formal, more friendly) or simplify the language. - When the agent wants to apply a custom instruction to the draft (e.g., "make more empathetic"). prerequisites :::info The Copilot appears in the ticket text box when the functionality is enabled on the account. If it doesn't appear for an agent, check the account settings with Cloud Humans support. ::: - Account with the AI Assistant feature enabled. - Agent logged in with permission to respond to conversations. - Internet connection (each generation makes a call to the AI service). About This Article Copilot is the AI assistant in Cloud Chat that helps agents improve responses directly within the ticket text box. It offers 8 predefined actions and 1 custom, then applies the result directly to the draft — you review, adjust if desired, and send. Frequently Asked Questions What does Copilot do? Copilot takes the draft you wrote in the ticket text box, processes it via AI, and returns an adjusted version based on the selected action. You can accept (replace the draft) or discard. What actions are available? | Action | Purpose | | --- | --- | | Rephrase | Rewrite with different words while keeping the meaning | | Correct spelling and grammar | Clean errors without changing style | | Shorten | Reduce length while keeping the essentials | | Expand | Add details to the text | | Make friendly | Adjust the tone to be more informal and approachable | | Make formal | Adjust the tone to be more professional | | Simplify | Use simpler and more direct language | | Translate | Translate into the chosen language (or conversation language in auto mode) | | Custom instruction | Apply a free-form instruction written by the agent | How to access Copilot? 1. Write a draft in the ticket text box. 2. Click the Copilot icon next to the send field. 3. Choose an action from the menu. 4. For Translate or Custom instruction, an extra field appears (target language or instruction text). 5. Wait a few seconds — the draft is replaced by the generated version. Can I undo a suggestion? Yes. After applying a suggestion, a toast (small floating notification at the top) with the "Undo" option appears for a few seconds. Clicking it restores the previous draft. :::success Undo restores the exact draft that existed before generation — including in conversations with multiple Copilot exchanges in the same session. ::: Is there a usage limit? Yes, 20 generations per minute per agent. This limit protects the AI service’s capacity. A typical agent usage is 5 to 10 generations per minute; therefore, the limit is rarely reached during normal operation. If reached, a message appears asking to wait a moment and try again. Cloud Humans reserves the right to modify this limit according to its policies. When does "Copilot temporarily unavailable" appear? This message appears when the Copilot AI service cannot handle the request. Common causes: - Incomplete credential setup on the account — check with Cloud Humans support. - AI service unavailable — temporary instability; try again in a few minutes. - Capacity limit exceeded on the upstream AI service — a short wait resolves it. :::warning If the message persists for more than 10 minutes straight, contact Cloud Humans support with account details, time, and attempted action. ::: Does Copilot send messages automatically? No. It only fills in the draft in the ticket text box. Sending to the customer is still your responsibility when clicking Send. The agent is always ultimately responsible for the content sent. Is the client's message sent to AI? For most actions (rephrase, shorten, expand, tone adjustment, simplify, grammar correction, custom instruction), only the draft you wrote is processed. For Translate, recent conversation messages may also be used as context to select the appropriate language and tone — limited to the last 10 messages. Can I disable Copilot on an account? Yes. The Cloud Humans team can disable the AI Assistant feature in account settings. When Copilot is turned off, the icon disappears from the ticket text box for all agents on that account. Does Copilot work on all channels (WhatsApp, email, chat)? Yes. Copilot works on the ticket draft regardless of the conversation channel. For channels with rich formatting (like email), the generated text appears as plain text — you can manually add formatting before sending. Notes - Copilot is a writing support tool — the final responsibility for the content sent remains with the agent. - The service behind Copilot is ClaudIA, Cloud Humans' AI platform. - For improvement suggestions or bug reports, contact the Cloud Humans support team through official channels.

Last updated on Sep 02, 2026

How to Show or Hide Activity Messages in a Conversation

The activity messages are the gray logs that appear in a conversation's timeline. They indicate actions such as assignments, status changes, modifications of custom attributes, and resolutions made by agents or automations. In conversations with many changes, these messages can make it harder to read the history. Therefore, it is now possible to temporarily hide them with a single click. What are activity messages? They are the gray messages that record internal conversation events, such as: - Agent assignment (e.g., "John assigned this conversation to himself") - Conversation resolution or re-opening (e.g., "The conversation was resolved by Ian") - Changes to custom attributes (e.g., "John changed test from true to false") - Automation actions (e.g., "Automation System (rule1) added test") Below is an example of a conversation with visible activity messages: Conversation with visible activity messages How to hide activity messages 1. Open any conversation in CloudChat. 2. In the conversation action bar (top), locate the eye icon. It is to the left of the email thread icon. 3. Click on the icon. The activity messages will be hidden immediately. Hovering over the icon displays a tooltip indicating the available action: Tooltip showing Show activity messages With activities hidden, only actual messages (sent and received) are visible in the timeline: Conversation with hidden activity messages How to show again Simply click the same eye icon. The activity messages will reappear in the original chronological order. Frequently Asked Questions Is the preference saved? Yes. When you hide activity messages, this preference is saved to your profile and persists across conversations and sessions. You won’t need to click again each time you open a ticket. Are activity data still being recorded? Yes. The toggle only hides the display of the messages. All activities are still being logged normally. When you click to show again, all activities will be there. Does this affect other agents? No. The preference is individual — each agent controls their own view. Do I need any special permission? No. All agents have access to the activity toggle. Tip: This feature is especially useful in long conversations with many attribute changes. Hiding activities allows you to focus on the message exchange history with the customer.

Last updated on Sep 02, 2026

Message Translation

Overview What is the Message Translation feature? Message Translation allows agents and supervisors to translate any text message directly within CloudChat. The translation appears in a side panel, without altering the original message — the customer never sees it, and the conversation history remains intact. Who is this feature for? Any agent or user with access to a conversation can use translation. No additional setup is required — the feature is available globally for all CloudChat accounts. How to Use How do I access message translation? 1. During a conversation, hover over the message you want to translate. 2. Click on the three-dot icon (⋮) in the corner of the message. 3. In the context menu, click "Translate" (indicated by a translation icon). What happens after I click "Translate"? A side panel opens to the right of the conversation. First, it displays "Translating..." with a loading indicator. After a few seconds, the panel updates to show: - Original Content — the message text as sent, with all formatting preserved. - A divider line. - Translated Content — the translated text with formatting preserved. Can I translate messages sent by the agent as well? Yes. The translation works for both received messages (from the customer) and sent messages (by the agent or bot). This is especially useful when a supervisor needs to understand what was responded in a language they do not speak. Can I translate audio, image, or video messages? No. The "Translate" option only appears on text messages. Messages that are solely images, audio, video, or files (without accompanying text) do not show the option in the context menu. Language and Settings Into which language is the message translated? The message is translated into the language set in your CloudChat account. For example, if the account is set to Portuguese (Brazil) (pt_BR), all translations will be into pt_BR. If it’s set to English, translations will be into English. Can I manually choose the target language? No. The target language is determined automatically based on your account settings. Manual language selection during translation is not available at this time. Where do I set my account language? The account language is configured in the general settings of CloudChat. If you need to change it, go to your account settings or contact your organization’s administrator. Behavior and Technical Details Does translation modify the original message? No. The translation is only a viewing feature in a side panel. The original message in the chat remains exactly as sent. The customer never sees the translation, and no text is changed in the conversation history. Is the message formatting preserved in translation? Yes. The translation maintains the original message formatting — bold, italics, lists, links, and other elements appear in both the original and translated content within the panel. Do I need to wait if I translate the same message again? No. Once a message is translated, the result is cached. Clicking "Translate" again on the same message displays the translation instantly, without loading. What happens if the translation fails? If it fails, the side panel shows the message: "Unable to translate this content". If that occurs, try again in a few seconds. If the problem persists, contact support. What technology is used for translation? Translation is performed using artificial intelligence (OpenAI GPT), not rule-based or dictionary-based systems. This allows for more natural translations with contextual understanding. However, as with any automatic translation, rare inaccuracies may occur — especially with very regional slang, specific technical terms, or ambiguous phrases. Recommended Use Cases I am an agent and received a message in a language I don’t understand. What should I do? Hover over the message, click the three dots (⋮), and select "Translate". Within seconds, the side panel will display the translation into your account’s language. You can proceed with the conversation normally — the translation is for your reference only. I am a supervisor and want to review interactions in other languages. Use translation on both agent and customer messages to understand the full conversation. Since the feature covers both received and sent messages, you can follow the entire context without needing to speak the language. I handle clients in multiple languages. Do I need to configure anything? No. Translation works automatically for any source language. Just ensure your account language is set correctly (e.g., pt_BR), and all messages will be translated to that language, regardless of what language the customer used. Additional FAQs Do I need to enable translation on my account? No. The feature is available globally for all accounts. No additional configuration per account or agent is necessary. Can the customer see that I translated their message? No. The translation is completely invisible to the customer. It appears only in the agent’s side panel, with no notifications or conversation changes. Does translation work for bot messages (ClaudIA)? Yes. Messages sent by the bot can also be translated — they are treated as sent messages, just like those from a human agent. Is there a limit to how many times I can translate? No. There is no documented limit per agent or conversation. Use the feature whenever needed. Does translation work offline? No. Since translation is performed via cloud AI, an internet connection is required to use it.

Last updated on Sep 02, 2026

Quick Filters (Filter and Sort) — FAQ

Quick Filters (Filter and Sort) — FAQ Overview What are Quick Filters? Quick Filters are the most direct way to control which conversations appear in your list and in what order. They are accessible via the "Filter and sort" icon in the icon bar at the top of the conversation list (next to the search bar). Clicking on it, a compact dropdown shows two fields: Filter by (conversation status) and Sort by (sorting criteria). What is the difference between Quick Filters and Advanced Filters? Quick Filters allow filtering by a status and selecting an ordering — it’s the fastest way to refine your list daily. Advanced Filters (accessible via the icon next to it, with tooltip "Advanced filters") open a separate modal where you can combine multiple criteria simultaneously (e.g., filter by tag + team + period). Use Quick Filters for everyday workflow and Advanced Filters when you need more specific searches. Filter by (Status) What filtering options are available by status? - Open (default): Active conversations in progress. This filter is automatically applied when entering the Conversations screen. - Waiting for Reply: Open conversations where the customer sent the last message and is waiting for the agent to respond. - Resolved: Conversations that have been closed or resolved (by agent or automatically). - Pending: Conversations not yet assigned to an agent or awaiting an action by the system (e.g., bot processing). - Snoozed: Conversations the agent "snoozed" to resume later (in 1 hour, tomorrow, next week, etc.). - Proactive Campaign: Conversations proactively initiated by the company (outbound), not by the customer. - All: Displays all conversations regardless of status. What does the "Waiting for Reply" filter mean? It’s an intelligent filter that identifies open conversations where the customer is waiting for a response from the agent. In practice, it shows conversations that meet all these conditions: 1. The conversation status is open. 2. At least one message has been received from the customer. 3. The last message from the customer is more recent than the last message sent by the agent — or the agent has never responded. It’s the quickest way to prioritize conversations needing immediate attention. Why is "Waiting for Reply" different from other filters? Because it is the only virtual filter in the list. All other filters (Open, Resolved, Pending, etc.) correspond to a real status saved in the conversation. "Waiting for Reply" is not a status in itself — it combines the "open" status with an automatic analysis of who sent the last message. For this reason, it does not appear in Advanced Filters and is not available as a condition in Automations. Does the selected filter get saved? Yes. The chosen filter is saved in your user preferences. When you exit the Conversations screen and return later (or close and reopen CloudChat), the last selected filter will remain active. What happens when I change the filter? The conversation list reloads from the server with results updated for the chosen filter. Depending on the volume of conversations, there may be a brief loading time. Sort by What sorting options are available? - Last activity: Newest first: Conversations with the most recent activity appear at the top. This is the most used sorting method daily. - Last activity: Oldest first: Conversations with the oldest activity appear at the top. Useful to find "forgotten" conversations. - Created on: Newest first: Recently created conversations appear first. - Created on: Oldest first: Oldest conversations are at the top. - Priority: High first: Conversations marked as high/urgent priority appear at the top. - Priority: Low first: Lower-priority conversations appear at the top. - Pending response: Longest first: Conversations waiting longer for a response are on top. Ideal for SLA management. - Pending response: Shortest first: Conversations waiting the least amount of time are on top. Which sort order should I use to manage SLA? Use "Pending response: Longest first". This orders the conversations that have gone the longest without a reply to ensure no customer is left waiting beyond acceptable limits. Which sorting helps find forgotten conversations? Use "Last activity: Oldest first". Conversations with no recent activity will rise to the top, making it easier to identify any stalled interactions. Combination with Assignment Tabs How do Quick Filters work with the "My", "Unassigned", and "All" tabs? Quick Filters work together with the selected tab. The tabs control who is responsible for the conversation, while Quick Filters control which status and which sort order to apply. Both criteria are combined automatically. Practical examples: - "My" tab + "Waiting for Reply" filter → Shows only conversations assigned to you where the customer is waiting for your response. - "Unassigned" tab + "Pending" filter → Shows pending conversations not yet assigned to any agent. - "All" tab + "All" filter + "Priority: High first" sort → Full view of all team conversations, prioritized by urgency. Recommended Use Cases I am an agent. How do I prioritize my workday? 1. Select the "My" tab. 2. Apply the "Waiting for Reply" filter. 3. Sort by "Pending response: Longest first". This ensures you see first the conversations that have been waiting the longest for your reply — where your attention is most needed. I am a supervisor. How do I identify conversations no one has picked up? 1. Select the "Unassigned" tab. 2. Apply the "Open" or "Pending" filter. 3. Sort by "Created on: Oldest first". This quickly finds old conversations that haven't yet been assigned to any agent. How do I monitor outbound campaign conversations? 1. Select the "All" tab (or "My", if you want to see only yours). 2. Apply the "Proactive Campaign" filter. 3. Sort by "Last activity: Newest first". You will see all conversations proactively initiated by the team, with the most active ones at the top. How do I review snoozed (delayed) conversations? 1. Apply the "Snoozed" filter. 2. Sort by "Last activity: Oldest first". This shows your snoozed conversations, with the oldest on top — useful for checking if any needs manual reactivation. Additional FAQs Can I combine multiple statuses in Quick Filters? No. Quick Filters allow selecting only one status at a time. If you need to combine multiple criteria (e.g., open conversations with a specific tag), use Advanced Filters. Why can't I find "Waiting for Reply" in Advanced Filters? Because "Waiting for Reply" is exclusive to Quick Filters. Since it is a virtual status (calculated in real-time based on who sent the last message), it’s not available in Advanced Filters or Automation rules. Why can't I use "Waiting for Reply" as a condition in Automations? For the same reason: "Waiting for Reply" is not a real status stored in the database. It is a dynamic check performed at query time. Automations operate on real statuses (Open, Pending, Resolved, etc.). Are the filters and sorting saved per device or per account? They are saved in your user preferences (account). Logging into another device will carry your preferences with you. Does text search respect applied filters? The search bar works independently of Quick Filters. Searching by text scans all conversations regardless of the active status filter.

Last updated on Sep 02, 2026

How the Queue Updates Work (“My”, “Unassigned”, and “All”)

"My" Queue - The "My" queue operates in real time. - When a new message assigned to the agent arrives, it appears automatically, without waiting for refresh cycles. "Unassigned" and "All" Queues There are some ways to update these queues: - Refresh the browser page (reload button). - Click on the queue name: this reloads the entire list, functioning as a refresh button. - Enable auto-refresh, which automatically updates every 30 seconds. - The auto-refresh setting is individual: turning it on or off does not affect other users on the team. Where to enable auto-refresh? Behavior when scrolling the queue with auto-refresh enabled (scroll) If you scroll down the list to view tickets that are not at the top of the queue and auto-refresh is enabled: - At the next 30-second cycle, the queue will reload automatically. - You will be taken back to the top of the queue, losing your previous scroll position. When to disable auto-refresh? It is recommended to disable auto-refresh in situations where you need to maintain your position in the list without being automatically taken back to the top. Examples: - Scrolling down to view tickets not at the top. - Curating tickets, such as reviewing, organizing, or prioritizing conversations. - Analyzing messages calmly without automatic updates interrupting the flow. - Manually assigning tickets or navigating between multiple conversations in the queue. With auto-refresh disabled: - The queue will not reload automatically. - You can work from your current position without losing context. After finishing your review or curation, just reactivate auto-refresh so the queue resumes automatic updating. Demonstration Video See the full functionality of auto-refresh and practical examples in the video below: 👉 https://www.loom.com/share/59185361e46e4e50b6ee21c7fcd6ba29

Last updated on Sep 02, 2026

How to use the Filters panel in Cloud Chat conversations

When to use it - You want to find conversations by combining simple filters (period, inbox, status, resolution, reason) without opening the advanced filters modal. - You have a conversation ID at hand (from the helpdesk or from CloudChat) and you want to open it without building any filter. - You want to see how many conversations the filter will return before applying it. - You want to know where the advanced filter you already used has gone. About this article The "Filtros" (Filters) panel is new and opens from the filter icon at the top of the conversation list. It does not replace anything you already used: - The advanced filters modal is unchanged and is opened from the "Avançados" (Advanced) link in the panel header. Details in How to create and use conversation filters. - The "Filtrar e ordenar" (Filter and sort) dropdown (one status + one sorting) is still next to it. Details in How to use quick filters in conversations. Filter icon at the top of the conversation list, which opens the Filters panel How it works Where is the old filter? The panel header has the Avançados link. It opens the advanced filters modal you already know, with the same fields as always, including the free-text filters and the AI filters. Nothing was removed: the panel is a shortcut for the most common cases, and the advanced modal remains the complete tool. Filters panel open, with the Avançados link highlighted in the header The two sets do not add up to each other. When you apply a filter from the panel while an advanced filter is active, CloudChat asks for confirmation before replacing the advanced one. The same happens when you click "Limpar" (Clear). A single field to search for a filter or an ID The field at the top of the panel serves two purposes: - Search for a filter by name. Type part of the name (for example, "resol") and the list shows only the categories that match. - Paste a conversation ID. If the text is a CloudChat ID, the panel offers to search directly for that conversation. If the text does not match any filter, it is offered as a helpdesk ID, so alphanumeric IDs work as well. Search field with an ID pasted and the suggestion to search for a conversation by ID :::warning Searching by ID replaces all the other filters. While an ID is applied, the other categories stay disabled. Remove the ID from its chip to go back to filtering normally. ::: Categories expand in place Each category ("Caixa de entrada" (Inbox), "Responsável" (Assignee), "Tags", "Status", "Resolução" (Resolution), "Motivo da resolução" (Resolution reason) and others) opens inside the list itself, without leaving the screen. Up to five options are shown; if there are more, the "Ver mais" (See more) button opens the full list for that category, with its own search. Motivo da resolução category expanded in the list, with the Ver mais button Multiple-selection categories accept more than one option. Single-selection categories (such as Gen CX Score) use one option at a time, and "Qualquer" (Any) removes the restriction. Period shortcuts The "Período" (Period) category offers the "Hoje" (Today), "7 dias" (7 days), "30 dias" (30 days) and "Mês passado" (Last month) shortcuts. One click applies the range; clicking the same shortcut again unselects it. For a specific range, open the category and pick the start and end dates in the calendar. Resolution reasons with readable names Resolution reasons are shown with labels in Portuguese (for example, "Inatividade do cliente" / Customer inactivity, "Frustração detectada" / Frustration detected, "Resposta completa da ClaudIA" / Complete answer from ClaudIA). In the full list, the original system value appears as a caption, for whoever needs the exact reference. Counter and chips for the selections The footer shows how many conversations the current filter will return, updated with every selection, before you apply it. The "Aplicar" (Apply) button is always visible. The 30 dias shortcut and the Resolvida status selected, with the conversation counter and the Aplicar button in the footer Above the list, the button with the count of active filters opens the selections as chips. Each chip removes its own selection; the period chip clears both ends of the range at once. Active filters button opened, showing the selections as removable chips Notes - The filter icon shows a counter with the number of filters applied to the list. - Free-text filters (other than the ID) and the AI filters remain available only in Avançados. - For the one status + sorting flow: How to use quick filters in conversations. - To combine many criteria and AI filters: How to create and use conversation filters.

Last updated on Sep 09, 2026

Ticket Kanban view: group by status or by your own statuses

When to use it - You want to see tickets on a board (Kanban) instead of a list - You want columns with your own statuses (e.g.: Triage, Under review, Waiting on customer, Done) Where to find it On the "Tickets" screen, in the upper right corner of the list, there is a pair of buttons: "Lista" (List) and "Kanban". Click "Kanban". :::info The Kanban button only shows up when the list has no advanced filters and no saved folder applied, and outside the "Menções" (Mentions), "Participando" (Participating) and "Não atendidas" (Unattended) views. If you can't see it, clear the filters and go back to the default Tickets view. ::: The native statuses are fixed Ticket statuses cannot be renamed or extended: | Status | Use | | --- | --- | | "Aberto" (Open) | Ticket in progress | | "Pendente" (Pending) | On hold, waiting for an internal action | | "Adiado" (Snoozed) | With a date/condition for automatic reopening | | "Resolvido" (Resolved) | Closed | When you group the Kanban by "Status", the board shows exactly these four columns. :::warning There is no way to create custom ticket statuses. For stages of your own, use the custom attribute path below. ::: How to create your own columns The Kanban groups by a conversation custom attribute of the List type — each value in the list becomes a column. Step 1 — Create the attribute 1. Go to "Configurações → Atributos Personalizados" (Settings → Custom Attributes). 2. Click "Adicionar atributo personalizado" (Add custom attribute). 3. Fill in: - "Nome para exibição" (Display name) — the name that appears in the Kanban selector (e.g.: Etapa do atendimento) - "Aplica-se a" (Applies to) — select "Conversas" (Conversations) - "Tipo" (Type) — select "Lista" (List) - "Listar valores" (List values) — type each stage and press Enter (e.g.: Triagem, Em análise, Aguardando cliente, Concluído) 4. Click "Criar" (Create). :::warning The type has to be "Lista" (List) — attributes of the "Lista Múltipla" (Multiple List) type do not appear in the Kanban selector. And it has to apply to "Conversas" (Conversations): "Contato" (Contact) attributes are not offered. ::: Step 2 — Group the board by it In the Kanban view, in the "Agrupar por" (Group by) field, switch from "Status" to the name of your attribute. The columns then become the values of your list, in the order in which they were registered. Step 3 — Classify the tickets The value is filled in on the ticket, in the conversation attributes panel (right sidebar). You can also fill it in automatically through Automations and Macros. Limits and behaviors - Column order — follows the order of the values in the attribute. To reorder, reorder the values in Settings → Custom Attributes. - "Sem valor" (No value) column — tickets without the value filled in fall into an extra column, which only appears when there is something in it. - Up to 20 columns — values beyond that do not become columns. - 25 tickets per column at a time, with "Carregar mais" (Load more) at the bottom. The counter at the top shows the real total. - Global filters (inbox, team, assignee, tags) apply to all columns. :::warning The Kanban is read-only: you cannot drag a card between columns or change the status from the card. Click the card to open the ticket and change it there. :::

Last updated on Sep 09, 2026

Email Sending Error Filter: The Three States and How to Use as a Work List

When to Use - You use the sending error filter to find contacts who haven't received a response, and it shows cases that the team has already resolved. - You want a reliable work list: only those who are still without a delivered response. - You want to understand why a conversation keeps appearing in the filter even after someone has interacted with it. How It Works The Three States The filter has evolved from a simple on/off switch to three options. Failure in the last send. This is the work list. It shows conversations where a sending failure occurred and no response has been received by the contact since then. Success in the last send. Conversations that had a failure at some point but where a subsequent response was delivered. Useful for checking what has been resolved. All. Any conversation with a history of sending failure, whether resolved or not. This is the behavior the filter had before. What Removes a Conversation from the Failure List Only one thing: a response that actually reached the contact. This means that two types of messages do not remove the conversation from the list: - Private note. It is internal and the contact never receives it. - Template, such as a satisfaction survey. It leaves the platform but is not the response the customer is expecting. :::info This distinction is the main point of the change. Without it, an internal note or a triggered CSAT could hide from the team a contact that still has no response. ::: The Filter Now Looks Only at Email The filter now considers only conversations from email inboxes. Previously, when using the "not equal" operator, it also returned conversations from WhatsApp and the widget, which do not have this type of delivery failure. Notes - If your operation used the old filter as a work list, switch to Failure in the last send. It is the equivalent state, excluding already handled cases. - A conversation may appear in Success in the last send but still need attention: the state indicates that the last response was delivered, not that the issue was resolved.

Last updated on Sep 22, 2026

Response Suggestion Copilot: ClaudIA suggests the reply for the agent

https://cloudchat.cloudhumans.com/rails/active_storage/blobs/redirect/eyJfcmFpbHMiOnsibWVzc2FnZSI6IkJBaHBBeTlBT0E9PSIsImV4cCI6bnVsbCwicHVyIjoiYmxvbl9pZCJ9fQ==--996f91cbd8d8565f2e30f3aa044c66c793f9a68f/copiloto-sugestao-externo.mp4 When to use - When the agent wants a draft response prepared, written from ClaudIA's knowledge base, instead of starting from scratch. - When the conversation has moved from ClaudIA to a human and the agent needs to respond quickly. - When operations want agents to respond with the same content used by ClaudIA. Prerequisites :::info The response suggestion copilot is activated by the Cloud Humans team. If it does not appear for your team, request activation through support. ::: - Cloud Chat account with the response suggestion copilot enabled. - Inbox connected to a ClaudIA project with a knowledge base. - An open conversation assigned to you. About this article This article explains how the copilot suggests responses to the human agent. It is different from the Chatbox Copilot, which improves, corrects, or translates a message you have already written. How it works What is the response suggestion copilot? ClaudIA reads the conversation, searches in the project’s knowledge base, and writes a response suggestion for the last customer message. The suggestion appears only for you, just above the text box. Nothing is sent to the customer without you clicking send. How to request a suggestion? There are two modes. Cloud Humans sets which mode applies to your account upon activation. | Mode | How it works | | --- | --- | | On demand | You click Suggest response whenever you want a suggestion. | | Automatic | The suggestion is generated automatically when the conversation is assigned to you and with each new message from the customer while the conversation is open on your screen. | What appears on the screen? While ClaudIA is writing, it shows Writing…. When it finishes, it displays Suggested response with the beginning of the text. Click to see the full suggestion. | Button | What it does | | --- | --- | | Use this | Places the suggestion into the text box. If you already typed something, the button changes to Replace what I wrote. | | Insert at cursor | Appears when you have already written something. Inserts the suggestion where the cursor is, without deleting your text. | | Try again | Requests another suggestion for the same message. | | Discard | Hides the suggestion. | After placing the suggestion in the text box, review, adjust if needed, and send as usual. What do the labels mean? - Low confidence: ClaudIA found little content in the knowledge base to support the response. Review carefully before sending. - Truncated: The response exceeded the maximum length and cannot be used. Click Try again. Can I ask for another suggestion? Yes. Click Try again. Up to 3 attempts per customer message are allowed, with a 30-second interval between them. In what language does the suggestion come? In the language registered in the contact. If the contact has no language set, the suggestion appears in the project’s language. Why did the suggestion not appear? - The conversation is not assigned to you. Only the assigned agent sees the suggestion. - The conversation is not open (e.g., already resolved). - The last message is not from the customer or is too short (e.g., "ok"). - You are in Private note mode. Switch to Respond to use the suggestion. - The project has reached the daily limit of suggestions. - You requested many suggestions in the last hour. Remarks - The suggestion uses the same content from ClaudIA’s knowledge base. If it was wrong or incomplete, update the knowledge base. - The daily limit applies to the entire project and is set by Cloud Humans. It resets at midnight. - To improve, correct, or translate a message you wrote yourself, use the Chatbox Copilot.

Last updated on Oct 07, 2026

Parallel Conversations: Trigger Another Department Without Leaving the Support

https://www.loom.com/share/cfde8d6333a2460d83e488035d5d98b7 When to Use - The agent needs the finance, legal, back office, or any other department to resolve the customer's case. - You want the request to this department to have status and owner — it’s not enough for someone to respond; the department needs to resolve it. - The agreement with the other department must be recorded within the support ticket, not in a loose email or message group. - The department you need to trigger does not use Cloud Chat and can only be reached via email. The end customer never sees a parallel conversation. Prerequisites The feature is active on your account, but neither channel works before being configured: - Internal sub-ticket — you need to create at least one inbox of type Parallel Conversation, one for each department to be triggered. - Email — you need to choose which email inbox the triggers will be sent from. :::info Only administrators can perform this configuration. The step-by-step instructions are in the Configuration (administrators only) section below. ::: How It Works What is a parallel conversation It is a separate conversation, created from a support ticket, to communicate with another department within the company. It has its own status, appears linked to the original ticket, and is invisible to the customer. As long as a parallel conversation remains open, the original support ticket cannot be resolved — the message displayed is "Resolve parallel conversations first", with the number of each. The two channels | | Internal sub-ticket | Email | |---|---|---| | Recipient | Departments using Cloud Chat | Departments outside the platform | | Where response is made | In their own queue | Responding to the email | | Who marks as resolved | The triggered department | You | Who sees what - End customer: nothing. - You, who opened: sees the conversation with the Requester View — message exchanges and response field. You do not enter the department's queue nor see their other tickets. - Triggered department: sees the full conversation, with status, responsible person, and history. Step-by-step Step 1 — Open the panel In the support ticket, on the right panel, open the Parallel Conversations section. It shows the count in parentheses if any exist. If none, it displays "No parallel conversations yet." Step 2 — Create the trigger Click + Create New. Choose the channel: Internal sub-ticket 1. In Recipient, select the department's inbox. 2. In Person (optional), choose a specific individual — or leave blank for the request to go to the team queue. 3. In Request, write what you need. 4. The AI-generated summary appears ready. Review, regenerate, or discard. 5. Use Attach to include files and click Send. Email 1. In To, search for a contact or enter the email address. 2. In Request, write what you need. 3. If you want to include the AI summary, check Include AI summary in email. It is unchecked by default, as the summary will be inside the email body and may contain customer data. 4. Use Attach and click Send. :::warning In the email channel, verify the address before sending. The request may contain customer data, and once sent, it exits the platform. ::: Step 3 — Monitor The parallel conversation appears in the panel with the Waiting label. When the other department responds: - an orange dot appears on the card, indicating New response; - a record is added to the original ticket's history. Click the card to open the conversation and reply. Back-and-forth exchanges are unlimited across both channels. Step 4 — Close - Internal sub-ticket: the triggered department marks the conversation as resolved. - Email: you read the response and mark as resolved — email respondents cannot change the status. The label changes to Resolved, and the original ticket is free to be closed. If the response did not resolve the issue, simply reopen the parallel conversation. Configuration (administrators only) Create a department inbox In Settings → Inboxes → Add inbox, select Parallel Conversation, name the department, and add its agents. Each department to be triggered needs its own inbox. Enable email channel In Account Settings → Parallel Conversations, under Email, choose which email inbox the triggers will be sent from. The default is None — until set, the email channel is unavailable to agents. Allow closing the original ticket In each parallel conversation inbox settings, there is an option Allow resolving the original conversation with open parallels. It is off by default — meaning the original ticket remains locked until all parallels are resolved. SLA pause While there are pending parallel conversations, the SLA of the original ticket is paused and resumes when the last is resolved. The history records both events. The first response is an exception: if the agent has not yet responded to the customer, this timer continues. To disable, go to Settings → SLA and adjust Pause SLA while waiting for parallel conversations in the desired policy. Notes - How the department knows it was triggered. If you assign a person, they receive a notification in Cloud Chat. If you assign only to the team, the request enters the queue without individual notification — someone must monitor the queue. No email or browser notification is sent. - Where to find parallel conversations. Only in the Parallel Conversations item in the sidebar menu. They do not appear in regular conversation lists. - Customer satisfaction surveys, automatic resolution, and follow-up are never triggered in a parallel conversation. - Reports. Parallel conversations are included in reports by inbox. To exclude internal work from metrics, remove parallel inboxes from filters. - Customer data identification. Internal parallel conversations do not load customer data. Context is provided by what you write in the request and the AI summary — describe the case clearly. - There is no limit to the number of parallel conversations per ticket.

Last updated on Oct 13, 2026

Act on All Conversations in a Filter at Once

When to Use - You need to apply the same tag to hundreds or thousands of conversations at once. - You want to resolve an entire backlog queue in bulk. - You want to assign all conversations from a filter to a team or a person. Prerequisites - The release is staged, carried out by our team, one instance at a time. If you don't see the option in your account yet, contact us and we will evaluate the activation. How It Works Select by Filter, Not by Scrolling Previously, bulk selection only included what was loaded on the screen: you had to scroll to the bottom of the list for each conversation to be included in the selection. Now, you set the filter and select all conversations it returns, even those that were never loaded. No Quantity Limit The action no longer has a limit on the number of conversations. A batch of thousands is accepted and processed in a controlled manner, in chunks, so as not to interfere with delivering responses to your customers. Monitor and Stop A banner shows the progress of the batch. If you need to interrupt, click Stop: processing ends at the boundary of the next chunk, without leaving conversations half-done. What has already been applied remains applied. The progress persists after refreshing the page and closing the tab, because the batch is registered on the platform, not in your browser. Unselect Before Applying After selecting everything by filter, you can unselect specific conversations. The counter updates, and the unselected conversations are excluded from the action. Step-by-Step 1. In the conversation list, set the filter with the segment you want. 2. Select all conversations from the filter. 3. Unselect those that should not be included, if applicable. 4. Choose the action: apply tag, resolve, or assign. 5. Confirm. The progress banner appears, and you can continue working. Notes - The Stop button appears for the user who initiated the batch, in the browser where it was started. - Large batches take a few minutes. Since progress is registered on the platform, there's no need to keep the tab open. - If another batch is already running, the platform notifies which one it is and its current status, instead of starting both at the same time.

Last updated on Oct 13, 2026

Revised Chat Area and Full Screen Image Viewer

The Cloud Chat conversation area has undergone a general visual review, and the image viewer has been redesigned: images now open occupying nearly the entire screen, with zoom and drag functionalities. When to Use - To read a receipt, screenshot, or document sent by the client in detail without downloading the file. - In daily customer service interactions: the visual review applies to the entire conversation area and does not need to be activated. Prerequisites None. The change is already implemented across all accounts and available to all agents. How It Works Full Screen Image Click on an image within the conversation. The viewer opens occupying almost the entire screen, and the image fills the available space. With zoom enabled: - Drag the image with your mouse or finger to reach any part of it. Dragging respects the edges so the image does not disappear from the screen. - Using the mouse wheel and click zoom in towards the cursor position, not from the center. Point to the area you want to read and magnify there. - Releasing drag outside the image does not close the viewer. - Switching attachments resets zoom and position to the starting point. Video and audio continue to work as before, just with more space. Visual Review of the Conversation Area The review aligned the list of conversations and the open conversation with the new design, without changing behavior. Visual changes include: - Conversation list cards, List/Kanban toggle, and bulk action menus. - Conversation header, Resolve and Reopen buttons, a group of more actions, and the merge modal. - Speech bubbles for text messages, images, contacts, locations, emails, including quoted emails and story replies. - Conversation banners: 24-hour window, parallel conversations, and auto-assignment. - Contact Details Panel: header, responsible person selectors, team, priority, attributes, tags menu, macros, and previous conversations. - Tooltips, loaders, and message scrollbar, which now only appears when hovering over. Notes - There are no configurations, permissions, or releases involved: the change is already active. - The screen behaviors remain the same. If something stopped working as before, contact us.

Last updated on Oct 20, 2026