Google reveals Pics, an AI image generator integrated into Workspace, aimed at competing with Canva in precise editing.

      Pics, powered by Nano Banana 2, allows users to create images from a text prompt and manipulate individual elements—such as moving, resizing, or translating—without needing to regenerate the entire composition. This feature is set to roll out to Workspace Business Standard and higher plans, as well as Google AI Pro and Ultra subscribers in the coming months.

      Google introduced Pics at the I/O 2026 developer conference on Tuesday as a new AI image generator and editor integrated within Google Workspace. The product aims to compete with design tools like Canva and Adobe Express, offering precision-editing capabilities designed to improve upon the 'prompt-and-pray' approach common in earlier AI image generators.

      Pics utilizes Nano Banana 2, Google's image model that the company claims is ideal for the application due to its ability to support accurate text rendering and detailed visual output. The combination of generating and editing allows users to alter, resize, and modify individual components within a design, adjust embedded text, and update specific areas of an image without needing to re-generate the entire graphic. This capability, termed localized object editing, is heavily marketed as a differentiator against single-prompt competitors.

      Another key aspect of the launch is the integration with Workspace. Pics will be directly embedded in Google Slides for in-deck image creation and editing, and users can save their work directly to Google Drive for sharing purposes. However, Google has not yet confirmed if Docs will be included, as the product page specifically mentions Slides and Drive while suggesting broader Workspace integration.

      Initially, Pics will be available through Google’s Workspace Experiments program for a limited group of early access testers; administrators within organizations using Workspace can opt-in by activating Gemini Alpha features. The full launch for Workspace customers on Business Standard and higher plans, as well as Google AI Pro and Ultra subscribers, is anticipated in the coming months. The company has not provided separate pricing details for these existing tiers, nor indicated if there will be an additional cost for Pics.

      Pics expands the Nano Banana family. Over the past year, Google has introduced the Gemini app, with Nano Banana 2 designed for precision tasks like text-on-image rendering and maintaining brand consistency. A product walkthrough by PetaPixel notes the difference between Pics and OpenAI's image generation in ChatGPT, framing it as a gap in workflow rather than model capabilities: ChatGPT is faster, while Pics provides more direct control over user edits post-initial generation.

      AI design tools have turned into a crucial area of competition among major model labs, with Pics being the most significant commitment made towards an integrated productivity suite as opposed to a standalone consumer application.

      Regarding provenance, the launch aligns with Google's SynthID watermarking feature, which has been standard on the company's generative-image outputs since 2023. OpenAI also adopted this earlier this year under the C2PA standard. While Google has not released a distinct provenance statement for outputs from Pics, the integration with Workspace suggests that the same SynthID watermarking applies as seen in the broader Gemini suite. DeepMind has consistently released watermarking as a default feature in its generative models through 2025 and 2026.

      What remains unclear from Google is the timeline for enabling this feature at the tenancy level beyond the 'coming months' phrasing, whether Pics will come at no additional cost for users under Business Standard, Business Plus, Enterprise, AI Pro, and AI Ultra tiers, any usage limits, the specific Nano Banana 2 model card and benchmark scores, the underlying compute requirements, and if the editing layer will eventually be available in Docs and Gmail.

      The Workspace Discord community and the Experiments newsletter are positioned as the primary channels for updating the public on staged availability. The next anticipated development will be the first general availability for AI Ultra subscribers and Business Standard customers, expected this summer, according to timelines outlined in third-party briefings.