Canva AI course · Social & promo
Social posts at pace
In the ChatGPT course you built the caption batch — a month of words from real moments. This lesson builds its missing half: the visual batch, produced at the same pace, in the same sitting rhythm, so posts ship as finished pairs instead of captions waiting on pictures. The two batches are a pair, not a repeat: words at the generalist desk, visuals here, married at the end.
Batch from the same moment list. The input is literally the same list your caption batch used — the month's real moments, gritty and specific. For each: does an honest photo exist? Real photo → the enhance path (background cleanup, crop, brand frame). No photo → the generate path, with your saved visual brief and the moment's specifics as the subject slot. The honesty line from the brand lesson routes this fork: product moments ride photos; concept moments (a tip, an announcement, a behind-the-scenes idea) may generate.
Templates carry the pace. The speed doesn't come from generating faster — it comes from deciding less: your post templates (photo frame, tip card, announcement layout) hold the design decisions, the brand kit holds the look, and each moment just picks its template and fills. Six posts in an hour is a template economy, not a talent surge — the same arithmetic as every frozen decision in this program.
Pair, then judge the pair. Marry each visual to its caption and judge them together: does the image show what the words claim? Do they repeat each other (waste) or complete each other (the caption adds what the image can't say)? The pair test kills the classic small-business post — a nice photo captioned with the same fact the photo already shows — and replaces it with pairs where each half earns its place.
Format physics live in the template. Platform rules — safe zones, where interfaces crop, how text sizes on a phone — get encoded once, in the template, exactly like caption format rules did in your prompt templates. The resize intelligence from the tour then propagates each post to its other formats. When a platform changes its physics, that's one template edit — the update-the-template rule, visual edition.
The thirty-second authorship pass, visual edition. Before anything ships, per pair: is the moment still true, does the image actually look like your feed's neighbor posts (lineup glance), is any generated element implying something real that isn't? Batch drafts are inventory here too — the pre-post glance is where you stay the author, and it's the same thirty seconds you already spend on the caption side.
One mistake to skip: letting the generate path eat the photo path. Generated visuals are frictionless; photos require walking to the shelf — and a feed that slides toward all-generated slides toward anybody's feed. For a business with a physical product or a real place, photos are the moat: imperfect light and true texture read as real in a feed filling with synthetic gloss. Generate the concepts; photograph the life.
Try it now (this lesson's task — social visuals): take your caption batch's moment list and run the visual half: fork each moment photo/generate, produce all of them through your templates, then marry and judge each pair. Ship the best two today. The pace you just felt is the lesson.