Video: "New Google Gemini Update is INSANE!" by Julian Goldie on YouTube.
Gemini Omni: what the new model actually does with video
Gemini Omni is the part of Google I/O 2026 that gets the most breathless coverage, so it's worth being precise about what it does. Omni is a multimodal model — it accepts image, audio, video and text input, and can output video grounded in real-world knowledge. That's different from Veo, which generates video from text prompts. Omni can take an existing clip, understand what's in it, and produce edited or extended versions with contextual awareness.
In practice, this means you can describe what you want changed, rather than re-recording it. For businesses already using video — product demos, how-to content, testimonials — the ability to repurpose and reformat without going back to a studio is genuinely useful. That said, Omni is in early access and the quality ceiling for complex production is still well below what a properly done shoot delivers. Use it where speed and iteration matter more than polish.
Veo 3.1 inside Google Vids: the editing story
Veo 3.1 is Google's video generation model, now integrated directly into Google Vids — their online video editor, part of Google Workspace. You work with a timeline, add text, images or audio, and Veo generates the visual layer to match. Compared with standalone tools like Sora or Runway, the Workspace integration is the selling point: everything stays inside the tools your team already uses.
For businesses that have avoided video because editing is a bottleneck, this is the more practical update of the two. It won't replace a video team on a brand campaign, but for regular short-form content — explainer clips, product spotlights, weekly updates — it reduces the gap between having an idea and having something publishable.
Ask YouTube: AI search inside YouTube, and what it means to be found
Ask YouTube is currently US-only for YouTube Premium subscribers, but the direction is significant. Instead of typing a keyword and scanning thumbnails, you ask a natural question — "how do I fix a leaky radiator" — and the AI pulls relevant video clips, structures a response, and lets you follow up conversationally.
From a content strategy perspective, this adds a new citation surface. YouTube already factors transcript content into how videos rank in standard search. Ask YouTube extends that — the AI reads transcript content to determine whether your video answers the question being asked. Clear spoken explanations, accurate auto-generated captions, and direct answers early in a video all help. It's the same principle as ranking in a Google AI Overview, applied to YouTube.
What this means now versus what to wait on
Not everything announced at Google I/O ships on the day. Gemini Omni and Ask YouTube are both in limited access. That's worth keeping in mind before restructuring a content workflow around them. The smarter use of this moment is to start thinking about what your video content library looks like to an AI model — whether your transcripts are accurate, whether your titles and descriptions are genuinely descriptive, and whether your videos actually answer the questions your customers are asking.
The tools for AI-powered creation are arriving in stages. The habits that make your content useful to AI models are worth building now, because they compound.
Where this connects to NordSys
Whether you're building a new site or improving an existing one, content that works for AI-first search has to be planned differently from content that just filled a blog. We build web pages that perform in both standard search and AI answers — with structure, clarity and genuine usefulness baked in from the start, rather than retrofitted later. If the Google I/O updates have made you think about how your site and video content look to an AI model, we're worth speaking to.
See our Web Design service →