Canva AI course · Visuals without a designer
The AI features that matter
Last desk on the map: the visual studio. Canva's proposition differs from every desk you've learned — its AI lives inside a design tool, so what you generate lands in an editable, brandable, exportable design instead of a chat window. This opening lesson tours only the AI features that matter for real work, and skips the rest. (Feature set current as of September 2026 — design tools ship AI features faster than any category, so the as-of date here is doing more work than usual.)
Generate-into-design. The core loop: describe a visual or a whole layout, get options placed on an actual canvas — a post, a slide, a flyer — where every element stays editable. This is the desk's defining advantage over standalone image generators: the output isn't a picture of a design, it's a design. You'll never screenshot-and-crop your way out of a generated image again.
Image tools that earn their keep. As of September 2026, the working set: text-to-image for when no photo exists; background removal (the quiet workhorse — product shots become usable assets in seconds); magic-style edits — erase an object, extend a photo's edges to fit a new format, swap a background. For a small business, the removal-and-extend pair alone replaces a monthly stack of fiddly editing.
Layout and resize intelligence. Feed one design, get every format: the square post becomes a story, a banner, a pin — with elements rearranged, not just squashed. Paired with brand kits (next-next lesson), this is where "a designer's afternoon" becomes "a minute per channel." The magic isn't creativity; it's consistency at speed, which is what a one-person business actually lacks.
Everything transfers, again. Describing what you want precisely (the five-part ask, visual edition — next lesson), iterating instead of regenerating from scratch, options over first drafts, judging with your own taste: the whole spine works here. The new muscle this course adds is visual specification — saying what you see in your head clearly enough that a machine can draft it.
Find your four locations. Same setup discipline as every desk, ten minutes: where generation lives (designs and images), where the brand kit is configured, where resize/ formats hide, and where export does (formats, transparency, print bleeds). Buttons will move — as of the date above, all four exist and are findable — and the re-find-after- redesign rule applies here more than anywhere.
Free tier and the upgrade wall, honestly. The pattern from every desk holds: start free, upgrade at a real wall — and on this desk the walls are concrete and findable: certain AI features are metered, some brand-kit depth and premium assets sit behind the paid tier, and export options vary. Run your first two weeks free with real deliverables; the feature you bump into three times is the one worth paying for, and you'll know its name by then. Paying to unlock a tool you haven't hit the walls of buys shelf-ware.
One mistake to skip: touring forever. Design tools are the most demo-able software there is, and an afternoon of "trying features" produces nothing but familiarity with menus. The counter-move you already know: bring one real deliverable to the first sitting — the post you owe, the price card for the fair — and let the tour happen inside making it. Features you used once for real beat features you've seen twice in a demo.
The working set, matched to the moments it exists for. Four situations from Maya's fair prep — match each to the tool that solves it.
Try it now: pick the one visual your work actually owes this week. In one sitting: generate three layout options for it, remove one background, resize the winner to a second format, export. Note which of the four locations you had to hunt for — that's your map, drawn by use.