Google has included campaign translation, seasonal imagery and video generation with Gemini Omni 1.2 in its second Creative Drop, listed in its announcements catalogue on 5 November 2026. The update brings together tools for preparing holiday advertising in Google Ads. Its focus is adapting existing materials to other languages, visual settings and screen formats, with a review of the resulting versions before they are used in campaigns.
The announcement concerns the production and localisation of advertising assets. Google describes three components: AI Campaign Translation for copy and campaigns, multimodal features in Asset Studio for imagery, and the updated video model for assets intended for YouTube and Google campaigns. Each component has its own workflow, while access to generation in Asset Studio requires the account to be eligible for generative AI features.
Translation creates a separate campaign ready for review
AI Campaign Translation supports adapting advertising into languages such as French, German, Spanish and Portuguese directly within Google Ads. For the localisation tool documented for Search campaigns, the source language is US English. Active and paused campaigns are eligible when their assets and keywords use the supported language. Removed campaigns, drafts and experiments are excluded from this workflow.
The process creates a new, independent campaign and preserves the original. It localises headlines, descriptions, keywords and elements such as sitelinks or callouts. For images with text overlays, the system can remove the text from the image and turn the translated message into separate text assets. This is an operation on the components of the advertisement, with those components reviewed within the new campaign.
The review interface shows the original text, its translation, a translation back into English and accompanying notes. Copy can be edited before publication. Google recommends examining the translated campaign while it is paused, and its documentation identifies that status as the default publication option. The team can therefore check terminology, brand names and the meaning of offers before enabling delivery to the selected audience.
Budgets and bids are copied from the original campaign, without currency conversion or exchange-rate calculations. Localisation also includes a landing-page component: advertisers can use manually updated addresses, domain replacement rules or dynamic translation through a JavaScript snippet provided by Google. These options determine the page a user reaches after clicking the advertisement, separately from translation of the advertising message itself.
The three components of Creative Drop #2
- Translated campaignsReview and publication in a paused state.
- Seasonal imageryProducts and brand guidelines as starting inputs.
- Video assetsConcepts, storyboards and formats for different screens.
Product photographs become the starting point for seasonal variations
The second component of Creative Drop #2 uses existing product images to build winter or gifting scenes in Asset Studio. Google presents uploading brand guidelines as a PDF and supplying several real product angles as control inputs. The brand's colours, fonts and wording rules can be provided in the same context as instructions about the desired scene.
The multimodal workflow accepts text, up to five product or asset-library images, and up to two PDF documents containing guidelines or creative briefs. English is the supported language for prompts and model understanding in the current documentation. That condition applies to generation within Asset Studio and is separate from the languages into which advertising campaigns can be localised.
A project can group images, copy and video concepts generated from the same requirements. Assets are organised by type and theme, then refined individually. Copy can be edited manually, images can receive different proportions or instructions, and preferred versions can be pinned to preserve them during subsequent AI changes. Producing a set of proposals and choosing the final assets are distinct stages of the interface.
Gemini Omni 1.2 joins the video workflow for YouTube
For video, Google has announced Gemini Omni 1.2, the updated version of Google DeepMind's video generation model. Asset Studio can create materials in horizontal, vertical and square formats. The company presents consistency and creation speed as improvements in this version. The model is used in Asset Studio's multimodal creation workflow, where video is built from the requirements and images supplied to the project.
Creating a standalone video begins with a creative source, such as an ad group or a landing page, supplemented with text and images. The system produces a brief, followed by a narrative proposal and a storyboard. Scenes, prompts and voiceover copy can be adjusted before the video is previewed. Uploading it to YouTube and including it in a campaign follow this preparation stage.
Google also proposes testing two different opening visuals during the first three seconds of an advertisement through the ad editor. Vertical and horizontal formats are presented together to accommodate different screens. Within multimodal projects, the system offers three video narrative concepts, one of which can be selected to generate the complete video. Concepts and completed videos are distinct stages of production.
Approved assets can be saved or added to campaigns
Asset Studio serves as a shared space for creation, editing, organisation and previewing. The account-level library gives account users access to saved materials, while previews can be sent to collaborators for feedback. Google also includes SynthID, its digital watermarking mechanism for identifying images generated or edited with AI. These features concern asset management, separately from budget settings and advertising delivery.
For Google Ads campaign management, the updates bring localisation, visual identity and video advertising preparation into the account's tools. In the workflow Google describes, completed assets can be sent to an existing campaign, used to prefill a new one or saved in the asset library. Generated project history is retained for 55 days, according to the Asset Studio documentation.
