Pollo AI says its guided creative tools use OpenAI models to help more than 26 million users produce images, videos, and advertising assets.
Pollo AI is using OpenAI models in a guided creation workflow designed to turn users’ ideas and visual references into images and videos. OpenAI published details of the collaboration on October 8, 2026, saying Pollo AI’s platform serves more than 26 million users creating videos, ads, and other visual content.
The new Pollo Agent combines models for different tasks rather than assigning every prompt to one model. Pollo AI uses GPT-5.6 to route requests and GPT-6 Astra for more demanding work, such as developing a narrative or planning scene revisions. The company selected GPT-Image-2.5 for image generation, including campaign visuals, posters, and storyboards. Pollo AI says this approach has reduced by more than 50% the time users spend choosing or switching between models.
The workflow is intended to handle creative direction as well as generation. It can develop storylines, scene sequences, and scripts from rough concepts and references. In a separate Marketing Studio feature, Photo to Video Ads turns a still product image and a prompt into a short social-media ad. OpenAI’s examples include a fragrance promotion made from a product photo and an approximately 15-second video, as well as campaign images for jewelry and a beverage brand.
Pollo AI says 30% of creator sessions already begin with its templates or trending-format workflows. The company’s broader aim, according to founder and CEO Bill Zhu, is to make AI video part of everyday content creation over the next 12 months. The examples show how the platform is positioning guided model selection and structured workflows as ways for creators and marketers to move from an initial concept to campaign-ready assets, while still relying on users to supply the creative direction.