SleekFlow Product Requests

Request: Provide an easy way for users to activate or assign an AI Agent to an ongoing conversation, without requiring the customer to send another message. Current Behavior When a conversation is already in progress: The customer sends a message. The conversation is currently being handled by a human agent. The user decides that the AI Agent should take over the conversation. The user adds the relevant AI enrollment label. The AI Agent is enrolled, but does not respond to the customer's latest message. The AI Agent only responds after the customer sends another message. This creates a gap where the customer's latest unanswered message remains pending even though the AI Agent has already been activated. Requested Behavior Allow users to assign/activate an AI Agent directly within an existing conversation. For example: Customer sends: "What are your opening hours?" Human agent reviews the conversation and decides the AI Agent can handle it. User selects "Assign to AI Agent" or "Activate AI Agent". The AI Agent immediately processes and responds to the customer's latest unanswered message. The customer does not need to send another message to trigger the AI. Why This Is Needed The current label-based enrollment creates an unnecessary dependency on the customer sending another message. This can cause: The customer's latest message to remain unanswered. Delays in customer response time. Human agents having to manually respond before the AI can take over. A poor customer experience, especially when the customer is expecting an immediate response. Confusion for users who assume that enrolling the AI Agent means it will immediately handle the conversation. Expected Outcome A user should be able to move an active conversation from a human agent to an AI Agent at any point, with the AI Agent immediately handling the latest unanswered customer message without requiring the customer to send an additional message.

For the AI Assistant response, I would like to have the AI Assistant to be able to know whats the date and time at that moment when drafting the message, it will help cross-checking with other contact properties and draft the best answer. E.g. Understanding the contact's birthday is coming up. In addition, I'd like the AI Assistant to be able to include URL when drafting the response to easily link the contact to the websites.

Feature request: Add Spanish and Portuguese language support to the AI Transfer and Summary handover. Current situation: When conversations are handled in Spanish or Portuguese, the AI Transfer and Summary features may still generate information in English, creating friction for local teams. Expected functionality: Support AI Transfer and Summary in Portuguese and Spanish. Generate summaries in the same language as the conversation when possible. Ensure AI-generated transfer notes and context are available in the selected language. Maintain the same functionality and quality across supported languages. Use case: Teams in Brazil and Spanish-speaking Latin America need to work in their local languages. Supporting Portuguese and Spanish would make AI-generated summaries and transfer information more useful and easier for local agents to understand and act on.

It would be great to have a feature that allows users to download or export all files uploaded to a Knowledge Base at once. Currently, there doesn't appear to be a way to download all original files at once that have been uploaded to the Knowledge Base. This can become a significant issue if the original file is accidentally lost or no longer available locally. In that situation, user can only retrieve the file one by one. It would be very useful to have an option to download/export in bulk: Download/export all files in the Knowledge Base at once, ideally as a ZIP file. This would provide users with a reliable way to back up their uploaded files, recover lost originals, make changes, save the time to download in bulk instead of download one by one.

We would like to request more control over Generated Articles in the Knowledge Base. At the moment, generated content may occasionally contain inaccurate information, and users can only edit it from the Article Wiki tab. Could we consider adding the following options? A Delete button for users to remove generated articles that are not accurate or no longer needed. A setting to disable automatic article generation. An option for users to review and approve generated articles before they are added to the Knowledge Base. These options would help users keep their Knowledge Base accurate and ensure that only approved information is used by the AI.

My team handles a lot of customer conversations over email, and those are the slowest replies to write. The AI can already suggest replies on chat channels, but when someone on my team opens an email thread there is no suggestion at all, so every reply gets written from scratch even when the answer is already in our knowledge base. I would like the same suggested replies on email. When a teammate opens an email conversation, the AI drafts a reply using the earlier messages in that thread and the knowledge the agent already works from, written the way an email should read rather than as a short chat message. My team still reviews and edits the draft before anything is sent, so a person always decides what goes out.

Automatically detect when your knowledge base articles contain contradicting information and flag it, so you know exactly what to fix. Resolving these conflicts helps your AI agent give more consistent, accurate answers.

This feature introduces a timeout branch for AI agent nodes, allowing the agent to automatically take action if a user does not respond within a set period. This ensures conversations remain active by enabling the AI to send follow-up reminders, helping to re-engage the user and move the interaction toward a resolution.

See how your AI agents are performing: how many active contacts they engaged, customer satisfaction on AI-handled conversations, and how quickly and how much they handled compared with your human team. Export the data for your own reporting.

This feature would be helpful for users who want to leverage AI analysis on business decisions when the company core language is not English. The feature can allow AI analysis to be translated according to the user's platform default language.