If a video can answer back, is it still a video?

D-ID's "Agentic Videos" put a real-time AI agent inside the viewing experience. Viewers can interrupt with voice or chat, ask about the subject, and keep the conversation going after playback ends. The launch release says responses are grounded in the original video script and can use additional connected sources. 1

Forbes describes the same shift as video becoming an interactive system rather than a fixed sequence. It also notes the unresolved boundary: grounding can reduce drift, but no system removes it entirely. 2

That changes the audience signal too. A question such as "Does this integrate with my CRM system?" is no longer a comment after the fact. It can become part of the media experience itself, alongside conversation volume, engagement time, question topics, and sentiment.

The question for Reddit: after the video ends, who owns the answer: the script, the connected sources, or the agent's interpretation?