Adobe supercharges Illustrator and Photoshop with new AI tools

If you spend half your workday nudging bezier curves, hunting down unidentified fonts, or warping flat artwork onto 3D packaging mockups, Adobe just shipped an update aimed directly at your headache.

Adobe released major feature drops across Illustrator and Photoshop. The updates blend their latest generative AI models into everyday workflows while finally fixing several long-standing interface annoyances.

Vector AI meets practical packaging tools

The marquee addition in Illustrator is the latest Firefly Vector Model in beta. It introduces Generative Shape Fill, allowing designers to sketch any custom silhouette and prompt Firefly to generate detailed vector artwork inside it, matched to the artwork's specific style.

Illustrator also rolled out Text to Pattern for generating seamless, scalable vector backgrounds, and a Style Reference tool that guides new vector subjects or icons to match existing brand aesthetics.

Yet the unglamorous utility tools might be the real crowd-pleasers. The new Dimension Tool lets designers plot exact measurements across packaging or apparel sketches in a few clicks. The Mockup feature automatically maps 2D designs across the curves and realistic lighting of 3D objects like bottles or boxes right inside Illustrator.

Need to edit flat text flattened inside an old image? Retype now identifies the exact font and converts static text into editable layers. Meanwhile, Illustrator's canvas pan and zoom engine received a performance boost up to 10 times faster, handling scaling from 3.13% all the way up to 64,000% without choking on dense files.

Photoshop speeds up targeted fixes

Over in Photoshop, Generate Image powered by the Firefly Image 3 model is now generally available. Generative Fill also gained an Enhance Detail toggle, which sharpens high-resolution assets so generative patches do not look muddy when printed or blown up.

Day-to-day manual editing got two new direct tools. The Selection Brush lets you paint selections onto the canvas rather than wrestling with complex lasso paths, and the Adjustment Brush allows quick, non-destructive tweaks painted directly over targeted areas.

Adobe also doubled down on its creator-first posture, emphasizing once again that it never trains on user files, relying only on Adobe Stock and public domain assets backed by Content Credentials.

Why it matters

We think Adobe is playing this right by refusing to treat AI as a standalone novelty. Instead of telling designers to prompt complete images from scratch, they are embedding machine learning directly into handles, paths, and brushes.

For working graphic artists and packaging pros, vector-native AI output is infinitely more valuable than raw pixel outputs from Midjourney, because vectors remain endlessly editable and production-ready. Features like instant mockups, quick dimension plotting, and font identification save tangible hours every week, making this update immediately worth adopting.

Competitors can generate stunning cinematic stills, but they still cannot slot cleanly into a print shop prepress workflow. Adobe's moat is not necessarily having the wilder imagination; it is owning the professional piping.

That said, Firefly's strict reliance on commercially cleared training sets can still produce outputs that feel slightly generic compared to unrestrained open models. But for enterprise teams terrified of copyright lawsuits, that clean provenance is precisely the feature that matters most.