What we are building next

A full breakdown of platform billing, with detailed interaction usage.

Build and edit AI agent workflows through a chat-style co-pilot in the workflow editor, with change indicators and undo.

Flag chat tags as high priority so those conversations jump the queue, either respecting or bypassing agent capacity limits.

Speech-to-text and text-to-speech for every language on the Proto platform, through an external API, with usage tracked per teamspace and a gender option.

Agents can add CC recipients to an Inbox conversation thread without creating a user profile, with automatic drop-off on removal.

Agents in the Inbox can call an AI assistant that drafts reply suggestions using the full context of the conversation.

A mediation mode gives co-pilot more room to produce a thorough, multi-paragraph reply for complex tickets.

The live report now shows each agent's day as a colour-coded 24-hour timeline instead of a single total.

Voice API now has its own prepaid billing set-up, with advance credit purchase, auto top-up, and usage broken down by request type, teamspace and language.

Restrict which domains can load your webchat, closing off unauthorised embedding of the script.

Deploy the AI agent as a standalone, full-page web experience, with its own setup, appearance and content settings.

Every change to a ticket or livechat is now logged as a timestamped entry in the conversation thread.

Maintain multiple saved versions of an AI agent's workflow and choose which is live, supporting testing and safe rollback.

Register external PostgreSQL or MySQL databases with secure credentials, SSL/TLS, and an optional SSH tunnel.

Render SQL results or structured JSON as a chart (bar, line, pie, and more), exportable as PNG/JPEG/SVG/CSV.

Run a SELECT query against any configured database connector at any point in a workflow, with success/error/timeout branches.

A single floating widget button surfaces multiple communication channels for the end user to choose from, with proactive messaging support.

A new auto fill field action classifies conversation content and auto-populates custom fields using AI, skipping low-confidence matches.

Email is now a dedicated channel, enabling AI-powered conversation handling over email.

AI agents can now call an LLM at any point in a workflow, not just as a fallback, with prompt variables, model choice, and success/error/timeout branches.

Overdue rules can now auto-reassign an overdue livechat or ticket to a live agent, one or more teams, or both.

Teams support alternative names for NLP-based auto-matching of incoming chats, with tags auto-applied on a match.

AI agents can now be added to Telegram group chats, working alongside live agents; the feedback form is not available in groups.

A gender option is now available for supported text-to-speech voices.

Custom voice model for speech-to-text voice recognition for Kinyarwanda.

Custom voice model for text-to-speech voice generation for Kinyarwanda.

Support for voice messages on the LINE channel.

Custom voice model for speech-to-text voice recognition for Oshiwambo.

Custom voice model for text-to-speech voice generation for Oshiwambo.

Roles and permissions can be used to grant or restrict access to specific Analytics metrics.

Summaries, sentiment and emotion analysis for chats and tickets.

Clear unread indicators and unread counts for conversations (chat and tickets) in the Inbox.

Improves the male and female voices for Tagalog.

Custom voice model for speech-to-text voice recognition for Tagalog.

Custom voice model for text-to-speech voice generation for Tagalog.