Agentic videos by D-ID - Interactive videos that talk back
by•
Turn any video into an interactive AI experience. With Agentic Videos, viewers don't just watch - they pause, ask questions, get real-time answers, and interact with the presenter inside the video itself. Viewers experience content in a fully personalized way. Creators gain a new world of insight into knowledge gaps and intent: a viewer who asks three questions tells you more than a thousand passive completions ever could. Now on the D-ID platform, built with industry-leading expressive avatars.
Replies
The insight about viewer questions revealing intent is really compelling. Can creators see a breakdown of which questions come up most, so they can improve future videos based on what people actually want to know?
D-ID
@doganakbulut Yes! Creators are getting an informative insights dashboard to see this, and many more conversation-based analytics.
This looks really useful — clean
execution. How does it handles
user interactions?
D-ID
@aditya_kalkotwar Viewers can ask free-form questions by voice or chat while watching, and the avatar responds in real time. The agent can be grounded not only in the video content itself but also in additional documents you can upload to its knowledge base, allowing it to answer questions with greater depth and accuracy.
Beyond spoken responses, the agent can present relevant visuals, including images and videos, to support its explanations, and it can interpret visual inputs as part of the conversation. Our goal is to make video an interactive, multimodal experience rather than a one-way medium.
D-ID
@aditya_kalkotwar Thanks! Under the hood, the viewer's question goes to an AI layer that's grounded in the video's content and any additional knowledge base you've connected. The avatar responds in real time and it feels like a natural continuation of the video, not a separate chatbot experience.
This solves something I've run into constantly with onboarding videos at work. Someone watches the whole thing, still has one specific question that wasn't covered, and either sends an email that takes days to get answered or just guesses and moves on. The idea of the video itself being able to handle that follow-up in real time is genuinely useful. Congrats on the launch!
D-ID
@anielka_cortes This is one of our motivators for creating this product! Having a question left unanswered, and then delays in response, creates a major pitfall in impact and retention. Allowing the viewer to work through the question in real-time reinforces the information in multiple ways.
Interesting concept but I'm skeptical about real-world adoption here. Most people don't naturally stop a video to interrogate it - the passive viewing habit is deeply ingrained. The creator insight angle is the most compelling part to me, but that only works if you get enough people to actually interact in the first place. Would love to see some retention and engagement data from early pilots before getting too excited. What are you seeing from actual users so far?
D-ID
@galdayan Great point, only outcomes count. We're indeed seeing that interaction currently happens mostly at the end of the video, supporting your point that passive viewing is a deeply ingrained habit. One of the key learnings at this point is that length, content depth, and narrative structure of the video determines when interactions naturally happen. While we are experimenting with interaction triggers, we're excited to learn more from our users about most valuable use cases.
Love the self service studio angle how much technical knowledge does a marketer or content creator actually need to go from a script to a finished Digital Person video?
D-ID
@amna9 One of the goals of the self-service studio is to make video creation accessible to non-technical users.
A marketer or content creator can start with a script, choose or create a Digital Person, select a voice, customize the scene, and generate a video directly in the platform. No coding or video production experience is required.
We'd love to hear what you think if you give it a try.
For teams already using video for sales enablement what's the typical turnaround time from uploading content to having a deployable Digital Person ready to embed?
D-ID
@ana_popescu2 If they're using a template, those are already available in the platform and can be created and embedded very quickly. If they’re looking for a custom v4 or any fully custom avatar, our team can typically build and add it to their account in about 10 days.
For uploading knowledge, the setup is fast and can be completed quickly. If they’re connecting via API, timing will depend on their internal dev team and the scope of their integration, but our API is built to make the process as smooth and efficient as possible.
Curious about the API is it built more for developers wanting full customization or can no code teams plug it into existing workflows without heavy engineering support?
D-ID
@andrew_paul11 Both!
How does D-ID approach voice cloning and lip sync accuracy across different languages? Is quality consistent for non English markets too?
D-ID
@antonio_manuel1 Great question. Multilingual content creation is a major focus for D-ID. The platform supports high-quality voice generation, voice cloning, translation, and lip-sync capabilities across a wide range of languages, and many customers use it to create content for global audiences.
Are there specific languages you're interested in? We'd be happy to share more details.
Given the 100FPS rendering capability are you targeting use cases like live streaming or gaming avatars as well or is the focus primarily on business communication and marketing content?
@carter_son Smart read on the rendering speed. The 100 FPS figure (120 FPS for enterprise streaming) is about processing throughput, not targeting live streaming or gaming as primary verticals. It means the rendering pipeline runs at multiple times real-time speed — which is what gives us sub-500ms conversational latency in live agent interactions. That's what makes the conversation feel natural rather than robotic.
Our current focus is business communication, sales enablement, L&D, and customer experience — that's where Agentic Videos and V4 Expressive Agents sit. The API is developer-accessible, so builders do create interesting things with it (we've seen EdTech, mobile, even game advertising — Gameloft used D-ID avatars for game character campaigns). But the platform isn't purpose-built for real-time in-game rendering or live broadcasting.
That said, I'm curious what you're building — would genuinely love to hear it.
This works excellent as an onboarding simulation!