August 2025

New Feature

Introducing Video Call Support with AI Avatars

Users can now enable live video call support directly within the Copilot platform, offering a more interactive and human experience.

• Comes with realistic visual avatars to align with your brand’s tone
• Great for product demos, onboarding sessions, and high-touch customer support
• Chat transitions seamlessly into a video call when triggered - ensuring uninterrupted conversations
• Supports full customization with avatar and voice selection across Web Widget, Public Webpage, and Mobile deployments

You can configure this under the AI Video Call section in your Copilot settings.

‍

New Feature

Attachment Support in Chat

Your Copilot agents now support file attachments, making conversations more interactive and informative.

• Upload images or PDFs directly into the chat
• Agents can analyze the content to answer questions, provide summaries, or extract specific details
• Easily enable this feature under Deploy > Surface > Attachments

This upgrade transforms passive file sharing into active, intelligent engagement.

New Feature

Users Can Now Capture Feedback with Quick Reactions

Your users can now give real-time feedback during conversations using simple 👍 or 👎 reactions.

• Users can rate individual chat messages from your AI Agents
• Feedback is automatically processed to generate conversation sentiment analysis
• Helps you identify patterns, coach agents effectively, and improve response quality over time

This feature enables a continuous feedback loop, making your AI Agents smarter and more responsive with every interaction.

New Feature

Add Real-Time Context with @ Mentions in Chat

Your AI Agent chat now supports @ mentions, making conversations more contextual, collaborative, and efficient.

• Use @ to tag contacts, teammates, or channels directly within the conversation
• Helps teams stay aligned in real time by adding the right context instantly
• Especially valuable for support, sales, and ops teams where fast collaboration matters most
• Developer note: Register a callback named "get_mentions_data" to enable mentions functionality
• This feature turns your AI Agent into a team-ready, context-aware assistant

📄 Learn more in the official documentation

New Feature

GPT-5 Models Integration

We’re excited to announce support for the latest GPT-5 model family—giving your AI Agents more flexibility, speed, and control over performance and cost.

• GPT-5 – High-performance model for advanced use cases, with deep reasoning and richer responses
• GPT-5 Mini – Lightweight, efficient, and ideal for balanced performance at low cost
• GPT-5 Nano – Ultra-fast and affordable, perfect for high-volume, real-time interactions

Whether you’re optimizing for power or scale, the GPT-5 family lets you tailor the intelligence behind your AI Agents to match your needs.