Google Vids adds conversational editing and instant AI avatars

Nobody wants to fix their lighting, clean their room, or put on a button-down shirt just to record a two-minute workplace update.

Google gets it. On Thursday, the company rolled out two major updates to Google Vids: conversational editing powered by Gemini Omni and personal digital avatars.

Directing AI with your voice

Back in February, Google added Veo 3.1 to Vids to handle text-to-video generation. But anyone who has generated video knows the pain: if the lighting is bad or the background looks off, you usually have to throw away the whole clip and roll the dice again.

Gemini Omni fixes that by introducing step-by-step conversational edits. Users can now generate clips using prompts paired with reference images like sketches or photos, and then edit them iteratively.

Whether it is an AI clip or raw footage shot on your phone, you can simply type instructions like "swap out the background" or "fix the lighting" without starting from scratch.

The clone that attends meetings for you

Then there is the personal avatar feature, which takes aim at on-camera fatigue. Users upload a single selfie and a short voice recording to create a digital twin that looks and sounds like them.

From there, you just paste your script. The avatar speaks the lines so you can deliver team announcements or training modules without ever hitting record.

Google is keeping guardrails tight. The feature is restricted to users 18 and older in select regions, locked specifically to the account holder's likeness, and stamped with invisible SynthID watermarks. Both updates are now available to Google AI Pro and Ultra subscribers alongside Workspace business tiers.

Why it matters

With OpenAI’s Sora shutting down, Google is proving that the real prize in AI video might not be Hollywood spectacle, but routine corporate communication.

By packaging conversational video editing and custom avatars directly inside Google Workspace, Google is landing a direct punch on startups like HeyGen, Synthesia, Captions, and D-ID. If you already pay for Workspace, paying another subscription just to generate a talking head video suddenly looks unnecessary.

We think Google’s decision to tie avatar creation strictly to authenticated account holders is brilliant risk management. It preempts the executive impersonation scandals that have plagued the category while giving enterprise IT teams peace of mind.

The real test will be social. When asynchronously watching five different AI clones give project updates on a Monday morning, teams will have to decide whether this actually saves time or just creates a strange new layer of corporate detachment.