Right now, Ask AI only accepts text input. But the model powering it (Claude Sonnet 4.6) natively supports image and file understanding — it can read documents, analyze photos, interpret code files, and work with uploaded content out of the box. That capability just isn't exposed in the current Ask AI interface.

Adding a simple attachment button (images, PDFs, code files, HTML, etc.) would unlock a massive amount of functionality that's already built into the model:

• Upload a screenshot of an email template and say "rebuild this in the email builder" — the AI can see it and replicate it

• Drop in a logo or brand asset and have it incorporated directly into whatever it's building

• Paste an HTML file or code snippet as an attachment for the AI to edit, debug, or improve

• Upload a PDF (like a service menu or price sheet) and ask the AI to pull info from it to build automations or content

• Share a photo of a before/after result and have the AI write social media copy or email content around it

The model already knows how to handle all of this — it just needs the interface to accept the files and pass them through. This would turn Ask AI from a text-only assistant into a true visual + document-aware builder, which is especially powerful for the email builder, website builder, and content creation tools.

This feels like a low-lift, high-impact feature since the underlying model already supports it. Just needs the upload UI and the file-to-API passthrough.