Feature Request

Business Problem: Users are limited by a 5,000-character restriction when defining persona instructions for the AI Agents. This might restricts their ability to provide detailed guidance, tone preferences, behavior instructions, and edge-case handling for the AI. Desired Outcome: Increase the character limit for instructions in the AI Agent to allow for more comprehensive customization. This would enable richer AI behavior and alignment with brand communication standards.

Business Problem: As a manager, I want to use AI to summarize calls after they end, so that I can quickly reference key points and action items without manually reviewing the conversation. Desired Outcome: Automatically generate concise summaries after calls. Highlight key decisions, action items, and follow-up points. Use Cases: Post-meeting review for client calls. Documentation of internal team discussions. The Current Workaround: N/A

We will be releasing email in 3 phases: Phase 1: Adding a one-click Gmail Workspace integration (only for Gsuite accounts). Status: BETA Released Phase 2: Adding a one-click Outlook integration (both personal and business accounts will work). See Status Here: https://respond.canny.io/feature-request/p/new-channel-email-outlook Phase 3: Adding IMAP/SMTP for Email to work with any third-party. See Status Here: https://respond.canny.io/feature-request/p/new-channel-email-imapsmpt --- Challenges --- > Structure of Emails & format < Emails are rather different from text messages. We aim to provide a unified experience for all channels, so we will have to make the experience of sending/receiving emails a lot simpler. > Execution of survey & automation < Most of the automated features provided by the respond.io platform are specifically designed for text messages. Many of these features wouldn't work well with Email out of the box. > Rate Limit for sending and receiving email < Different email clients have limits on the amount of outgoing email per day. If the limit is reached, users will not be able to answer emails on that channel until the next day.

Business Problem: As a first-point-of-contact agent, I want call transfer capabilities, so that I can quickly route calls to the appropriate team or individual, ensuring efficient handling of customer inquiries. Desired Outcome: Ability to transfer live calls seamlessly to another agent or department. Use Cases: Transferring a client to a sales representative for advanced queries. Transferring support calls to specialized teams during escalation. The Current Workaround: N/A

Business Problem: Currently, each workspace can only create up to 20 AI Agents , which can cause operational and performance issues: Hard for Owners and Managers to craft clear, well-structured prompts when different use cases have to be squeezed into fewer Agents. Forces long, mixed instructions into single agents, reducing accuracy and increasing irrelevant or off-scope replies. Combining multiple contexts in one agent makes it harder for the AI to follow guidance consistently and deliver reliable results. Desired Outcome: Increase the limit and allow workspaces to create more than 20 AI Agents to enable better specialization, clearer context separation, and improved response accuracy.

Business Problem: Today, the AI Agent can tell the current time, but it isn’t consistently reliable at deciding whether a given time falls inside or outside a specific window (like “Mon–Fri, 9am–5pm”). That leads to occasional misfires: the Agent may say you’re open when you’re closed (or vice versa), especially around timezones, weekends, holidays, or boundary times like exactly 5:00 pm. This creates inconsistent experiences and erodes trust. Desired Outcome: Give the AI Agent rock-solid, predictable time awareness. It should confidently determine whether now is within your business hours, understand special hours and holidays, and tell customers how long until you open or close next. It should also handle simple date checks (e.g., whether today falls within a promotion period) and basic “time until/from” questions. The result: accurate responses, fewer false “open/closed” moments, and a smoother customer experience every time.

Business Problem: Right now, the AI Agent only remembers up to 50 messages in a single chat and forgets everything once the conversation ends. When a customer comes back later or gets passed to another agent (human or AI), the AI loses context. This leads to repetitive questions and a poor experience. Desired Outcome: Ensure the AI can retain important context across longer conversations and multiple sessions. This allows it to deliver more consistent, relevant, and personalized responses — even when the chat is handed off or resumed later. As a result, users won’t have to repeat themselves, and conversations can continue smoothly. To achieve this, the AI should: Remember more than 50 messages in a single conversation. Keep key info when chats are reopened later. Continue conversations smoothly even when switching between agents. Use Case (All Lifecycle Phases) Capture : AI remembers earlier inputs, so users don’t have to start over if they come back later. Convert : Keeps track of past preferences or questions to guide better sales conversations. Retain : Remembers past issues or interactions to provide consistent and personalized support.

Business Problem: Currently, AI Agents can only process a limited range of file types — mainly text, images and files. However, many customers receive valuable information through audio messages. The inability to parse and respond to these content format restricts the AI Agent’s usefulness and limits automation potential. Use Case Pain Points: Audio files : Customer send voice notes or support queries via platforms like WhatsApp or Messenger. These aren’t transcribed or actionable by the AI Agent. Desired Outcome: Audio files : Transcribe and analyze content (e.g. .mp3, .wav)

Even when I explicitly instruct the AI Agent (via prompt) to avoid responding outside the defined scope, it continues to generate incorrect information. The agent is making promises to perform services instantly, providing business addresses that are not ours, and telling customers to contact emails that do not exist. I need a way to restrict the agent from inventing or hallucinating information.

We currently track the status of the user (active/inactive) to decide whether we should send an email notification or not, based on their interaction with the platform. This will now be improved into the User Activity Status feature. User ActivityStatus will grow to become an important part of many other functions. For example, the auto-assignment feature will have an option to only assign new contacts to users who are currently active. Users will be able to manually change their status on the platform between Online (Active), Away (Inactive), and Offline (Inactive) . The Status will also be changed automatically (e.g. based on inactivity) depending on the Space settings `` configurations. `` Future Use Cases / Potential Use Auto Assignment to only active users on the platform. Team members / users real-time overview. (show status or last signed in / hours clocked in)