Whisk is Google Labs' image remix tool that generates a new image from three reference photos: a subject, a scene, and a style. PR teams use Whisk to prototype campaign visuals and social concepts quickly, without writing a detailed text prompt or briefing a designer first. This tool enables rapid visual iteration for communication professionals.
What is Whisk?
Whisk is a Google Labs experiment that combines three uploaded images, one each for subject, scene, and style, into a new generated image. This process eliminates the need for a long written description. Google launched Whisk in the US in December 2024 and expanded it to more than 100 countries by February 2025, according to a report on the international rollout.
Whisk operates using two Google models: Gemini and Imagen 3. Gemini reads and describes the content of each uploaded image. Imagen 3 then generates the new picture based on those descriptions, as stated in Google's own launch announcement. The tool captures the essence of each reference image without reproducing it exactly.
How does Whisk differ from text-to-image tools like Nano Banana?
Whisk differs from text-to-image models, such as Nano Banana, because Whisk takes images as its primary input. Text-to-image models, however, begin with a written prompt.
Why it works: Whisk's three-slot structure for subject, scene, and style compels users to think in concrete visual references instead of adjectives, according to Google's product description. This specific mechanism helps keep the output closer to what a communications team envisions. A team with an approved brand photo, a location shot, and a mood reference can directly combine these elements. This approach avoids translating the combination into words that a text model must interpret.
How are PR teams using Whisk today?
PR teams are using Whisk to test visual directions before allocating budget for a photo shoot. They input a product shot, a target venue, and a style reference to generate several remix options quickly. These teams then present the strongest visual concepts for internal review or client presentations.
For example, a consumer brand communications team pitching three creative directions for a product launch can build each concept in Whisk. They use the same product photo, paired with different scene and style references for each direction. This allows them to present all three options side-by-side to leadership before booking any photographer. Whisk also includes default style presets, such as Sticker, Plushie, Capsule Toy, Enamel Pin, Chocolate Box, and Card, according to an independent product breakdown. These presets offer fast, on-brand starting points for social content without requiring a custom style reference.
What should PR teams check before using a Whisk image externally?
PR teams should verify that any Whisk output used publicly does not closely resemble a real, identifiable person from the reference photos. Whisk is designed to remix and reinterpret images, not reproduce them exactly. A close resemblance to a real face in a published asset could raise questions of likeness and consent.
Google notes that Whisk remains a Labs experiment, and its availability varies by country. For instance, the UK and India were unavailable as of the most recent international expansion, according to the same rollout report. Therefore, a global communications team must confirm access in each market before building a campaign workflow around Whisk.
When does a PR team still need a real photo shoot instead of Whisk?
A PR team still needs a real photo shoot for any asset requiring factual accuracy of an actual product, spokesperson, or venue. Whisk output is a remix and reinterpretation, not a faithful photographic record of uploaded content.
What is Whisk Animate, and why does it matter for PR?
Whisk Animate converts a Whisk-generated still image into a short video clip using Veo. This feature is available to Google One AI Premium subscribers, according to an independent product database that tracks product updates.
For a communications team, Whisk Animate extends the fast concepting workflow into short teaser clips. A still image approved in Whisk can become a first-pass animated version without switching tools or starting a separate video brief. This capability is useful for testing a social video direction before committing to full production.
| Use case |
Whisk capability applied |
Trade-off to plan for |
| Early creative-direction testing |
Combine a subject, scene, and style reference to preview several concepts fast |
Output reinterprets the references rather than reproducing them exactly |
| Social content in a house style |
Use built-in presets like Sticker or Enamel Pin for consistent, on-brand posts |
Presets read as playful, not appropriate for every brand voice |
| Client concept presentations |
Generate multiple visual directions from the same source photo |
Not a substitute for a final campaign asset needing factual accuracy |
Conclusion
Whisk's image-in, image-out approach solves a problem text-to-image tools do not: it allows a communications team to start with trusted references. This includes real product photos, locations, or mood boards, instead of translating judgment into a written prompt. This makes Whisk a fast method for testing creative direction before a shoot. Teams must limit the output to concepting and early review, not final, factually accurate assets.
