Google Vids Adds Personalized AI Avatars Powered by Gemini
Google is expanding its Vids product with AI-generated personal avatars that let users create videos starring a digital version of themselves, backed by Gemini Omni for prompt-driven video generation and editing.
Original sourceGoogle has updated its Vids platform — part of the Google Workspace suite — to allow users to create AI-generated videos that feature a digital avatar modeled on themselves. The feature lets users record a short reference clip, from which the system builds a personalized avatar that can then appear in AI-generated video content without requiring additional filming sessions.
Underpinning the avatar and video generation capabilities is Gemini Omni, Google's multimodal model, which powers both the prompt-to-video generation workflow and the editing tools. Users can describe a scene or a message in natural language and have Vids produce a video with their avatar delivering it, complete with background, pacing, and narration.
The move positions Google Vids directly against tools like Synthesia and HeyGen, which have built businesses around exactly this use case — personalized AI presenter videos for corporate training, sales outreach, and internal communications. Google's advantage is distribution: Vids sits inside Google Workspace, which is already deployed across millions of organizations, meaning the sales cycle for this feature is effectively zero for existing customers.
The feature is currently rolling out to Google Workspace subscribers, though availability varies by tier. No standalone pricing for the avatar feature has been announced separately — it appears bundled into existing Workspace plans, which could significantly undercut the pricing models of dedicated competitors.
Panel Takes
The Skeptic
Reality Check
“This is a direct shot at Synthesia and HeyGen, and the distribution advantage is real — Workspace is already in the building, no procurement needed. The question is output quality: enterprise avatar tools live or die on whether the avatar looks credible at 1x speed on a laptop screen during a sales call, not whether it demos well at a conference. If Gemini Omni's avatar fidelity is at Synthesia's 2024 quality level or better, this kills the mid-market competitor segment within 18 months. If it's noticeably worse, IT admins will hear complaints and route around it.”
The Creator
Content & Design
“The workflow problem this solves is real — re-recording a two-minute explainer video because you changed one sentence is genuinely painful, and avatar-based editing fixes that. But the taste layer here is entirely delegated to Gemini's defaults, and that's where these tools consistently fall apart: the avatar blinks on a rhythm, the background looks like a stock photo from 2019, and the delivery has the cadence of a phone tree. Until Google shows me a gallery of actual output that doesn't read as 'AI video,' I'm not convinced this produces anything a creator would choose to ship over a direct-to-camera phone recording.”
The Founder
Business & Market
“Synthesia raised over $150M to build exactly this distribution problem — getting into enterprise orgs — and Google just made that money largely irrelevant by bundling it into Workspace. The moat for dedicated avatar video tools was never the technology; it was seat-level adoption and workflow integration, and Google owns both channels. The specific companies I'd be worried about right now are the ones whose entire ARR sits in the 'corporate training video' use case with mid-market Workspace customers, because that renewal conversation just got a lot harder.”
The Futurist
Big Picture
“The thesis embedded in this feature is that synchronous, camera-on communication is becoming optional — that a credible digital proxy of you is sufficient for most internal business video, and that the friction of filming yourself will eventually feel as archaic as typing your own HTML. That bet requires two things to hold: avatar fidelity crossing an uncanny valley threshold users find acceptable at work, and organizational culture tolerating 'I sent my avatar to that all-hands.' The second dependency is underrated and culturally uneven — it will clear in some companies and stall completely in others, which means adoption curves will look geographic and sector-specific rather than smooth.”