Claudia - Feature Requests

Today, specialist agents execute sequentially within the same response shift. Request: evaluate support for parallel execution between expert agents, to reduce response time when multiple agents need to be consulted in the same interaction. Source: Unthread #19390 - Solides: https://cloudhumans.unthread.io/dashboard/inbox/all/conversations/7b8ef940-ad24-4a1a-8ae3-138773baba1b

In the Daki project, many cases are about “my order was canceled”. The chatbot that precedes the service can already read the order and provide information such as status (if canceled) and payment method, but CLAudia does not have access to this data when the ticket is processed. Request: give CLAudia access to this information (via direct reading or an agent that provides it) to allow more personalized answers, such as confirming the cancellation and the refund period automatically. Source: Unthread #19528 - Daki/Jokr: https://cloudhumans.unthread.io/dashboard/inbox/all/conversations/1601f039-91ab-428e-aa4f-fd302be09f92

When an instability occurs in the product that generates a peak of doubts on the same topic, today it is necessary to manually update, one by one, the contents of the database that address the subject. Request: a flow/automation (for example via subagents) to update in bulk the content related to a specific instability, and a way to map/measure in real time the volume of customers impacted by it. Source: Unthread #19230 - Simple Account: https://cloudhumans.unthread.io/dashboard/inbox/all/conversations/7f1b45ab-614f-4e92-9248-bbd12937fc9c

Connecting Cloud Humans' MCP to an AI assistant (Claude), the customer wanted to assemble a summarized daily update of the operation, but today the MCP (beta) does not show CSAT/satisfaction metrics or retention rate per channel. Request: make these metrics available via MCP to allow automated reporting by the client itself. Source: Unthread #19728 - ch-liquidz: https://cloudhumans.unthread.io/dashboard/inbox/all/conversations/ef28b4f0-3ea6-4591-8869-4a9e2d3ea189

Request to associate the contact reasons (Category, Reason, and Subreason) with the tags and FAQs in the knowledge base. This association does not exist today: there is no way to link a FAQ/content directly to the reason for contact classification used in the reports, which makes it difficult to correlate the content answered with the actual reason for the service. Source: Unthread #19312 - Online Shipping: https://cloudhumans.unthread.io/dashboard/inbox/all/conversations/429398ee-4856-4b30-b46f-f196002741fe

Request for behavior improvement from CLAudia: (1) automatically identify when another bot is responding in a loop (e.g., automatic messages from another service) and escalate to N2 in these cases; (2) do not reply to an email thread that has already been answered by a third party (e.g., partner like asksuite.com ), avoiding duplicate/unnecessary responses. Source: Unthread #19639 and #19644 - Asksuite: https://cloudhumans.unthread.io/dashboard/inbox/all/conversations/862a4a21-ee9a-4915-8fed-fc9ebaf0b24d

The visualization of data in the new dashboard worsened on specific points compared to the old one: the retention rate does not appear in some cases and is visually very close to the volume of tickets; the CSAT no longer shows the number of votes received (only the grade); and the new dashboard has subtitles in English that should be in Portuguese. The overall layout is cleaner, but these points worsened the reading of the data. Source: Unthread #19795 - Solides: https://cloudhumans.unthread.io/dashboard/inbox/all/conversations/1e71a6be-4840-49d8-9d04-bf500aa3fdeb

Today, to investigate why the agent was (or was not) triggered on a ticket and adjust the triggers, it is necessary to exit the conversation and go to Agents > AI companion > Investigate a conversation, pasting the ticket link there. Suggestion: bring this investigation/trigger adjustment shortcut to the conversation screen itself, avoiding navigation between screens.

When the chat widget is in the client's logged in area and their language is already known, it would be useful to be able to pass this information via widget/form so Claudia can already identify the correct language, avoiding Translator failures with short messages. Source: Unthread #8463 - Asksuite: https://cloudhumans.unthread.io/dashboard/inbox/all/conversations/090cd757-8b20-4d07-ae3e-fccf9ded2ed4

Today it is difficult to identify what is enabled or disabled in CLAudia's content compared to the Hub: before the gray text indicated disabled items, now there is no visual distinction and the items are separated into tabs. Additionally, enable/disable requires using the pencil icon, without a direct button. Source: Unthread #19171 - Chips: https://cloudhumans.unthread.io/dashboard/inbox/all/conversations/93acd693-7d6b-4184-8017-6842f21f9f99