Air launches 50 AI models, canvas editor, and usage-based pricing in biggest product release ever
Mar 23, 2026 with Shane Hegde
Key Points
- Air launches access to 50 AI models, a canvas editor for asset variants, and AI agents that execute edits autonomously, positioning the release as automation of mechanical work rather than creative replacement.
- The company introduces a brand context layer that stores organizational memory—fonts, guidelines, photography style—enabling non-creatives to generate on-brand assets while agents enforce consistency or reject off-brand edits.
- Air shifts to usage-based pricing across its product and argues the tech industry's "CMO killer" messaging is self-sabotaging, saying companies should automate specific jobs rather than market products as role replacements.
Summary
Air is launching its largest product release to date with access to 50 AI models, a canvas editor, AI agents for automated edits, a brand context layer, and a shift to usage-based pricing. The release ships tomorrow.
Shane Hegde, Air's co-founder, is positioning the launch against a narrative he calls dangerous: that AI will make creative roles obsolete. Air ran a New York Times ad campaign borrowing from Good Will Hunting to argue the point. AI can recite facts about art but cannot understand the texture of lived experience. While generative models are powerful, creative work remains deeply illogical, requiring extreme subjectivity and perfectionism in ways machines cannot replicate.
Model access and canvas editing
Air is integrating 50 AI models directly into its product. Users can grab existing assets and create variants by changing text, turning images into GIFs or video, or generating new versions through a visual canvas editor.
Agents for automated edits
Users can instruct agents to make edits on their behalf, shifting from hands-on editing to conversation with machines.
Brand context layer
The differentiator is a brand context layer that stores organizational memory within Air's system. It holds fonts, styling, photography style, and brand guidelines. Agents can apply this context to maintain consistency or decline edits that violate brand parameters. This lets non-creatives generate on-brand assets without direct creative oversight.
Usage-based pricing
Air is moving entirely to usage-based pricing across the product.
Hegde also emphasizes Air's existing strength in AI-powered search. The company has spent two years building natural language search powered by computer vision models, including facial recognition, scene detection, and object detection. This lets teams find assets by meaning instead of filename. The new release merges search with generative variant creation, so users can find assets semantically and then generate variants from them in a single flow.
The positioning is deliberate: rather than replace creatives, Air is automating away the mechanical parts of creative work to give creatives space for what requires taste and judgment. Machines handle execution and consistency. Humans direct strategy and vision.
Hegde's critique of the industry is direct. Tech companies keep marketing AI as a "CMO killer" or "creative killer," which is self-sabotaging. Telling customers they are irrelevant is not a viable go-to-market strategy. Companies should instead build products that meet humans where they are and automate specific jobs to be done, not entire roles.