← Video Library · Content & Video
Create Engaging Social Media Shorts with ChatGPT + Canva!
Published December 5, 2024 · 45 views on YouTube
Key takeaways
- Prompt ChatGPT to generate a CSV of content in columns (question, tip one, tip two, tip three) that map directly to a Canva template.
- Film one plain, silent template video of yourself giving a tip, with headings for question and each tip timed to your gestures.
- Use Canva's Bulk Create app (Canva Pro required) to connect each CSV column to a text layer, then generate dozens of finished videos at once.
- No on-camera footage needed — drop in a stock video from Canva's library as the background instead of yourself.
- Feed ChatGPT an existing blog post or presentation instead of a fresh prompt to repurpose material you already have into the same tip format.
How do you use ChatGPT and Canva together to mass-produce social media shorts?
Erica Hannah’s method: prompt ChatGPT to generate a CSV file of short-form content ideas in a fixed column structure (question, tip one, tip two, tip three), then use Canva’s Bulk Create app to connect those columns to text layers on one template video. Canva generates a finished, ready-to-post video for every row in the CSV — turning one template and one prompt into dozens of Shorts, Reels, or TikToks.
What do you prompt ChatGPT for, exactly?
Ask ChatGPT for video marketing tips in a conversational, easy-to-understand format, structured as: question, tip one, tip two, tip three, with each element in its own column, delivered as a CSV file. Erica requested 10 examples in her demo, but noted you could ask for 20, 30, or 50 depending on how much content you want in the batch. Download the resulting CSV once ChatGPT finishes generating it.
How do you build the template video in Canva?
- In Canva, create a new design and choose the format you need (e.g., “YouTube Short,” which auto-sizes the canvas).
- Upload or select a short video with no spoken audio — Erica pre-filmed herself waving, pointing, and gesturing through “tip one, tip two, tip three” beats, but a Canva stock video works as a silent background too.
- Add a text heading for “Question” and set its background/effect so it’s easy to read.
- Turn on “show element timing” to see exactly when each clip beat happens, then drag the text layer’s timing to line up with your gestures in the video.
- Repeat the same steps for each tip heading (Tip 1, Tip 2, Tip 3), matching each text layer’s timing to when you point or gesture for that tip in the footage.
How does Canva’s Bulk Create app turn one template into many videos?
- Open the App Store inside Canva and search for “Bulk Create.”
- Select “Upload data” and upload the CSV file downloaded from ChatGPT.
- For each text layer in the template (Question, Tip 1, Tip 2, etc.), click “Connect data” and link it to the matching CSV column. Skip any column you don’t have a layer for (Erica skipped Tip 3 in her demo since her clip only covered two tips).
- Click “Continue,” then “Generate [X] designs.” Canva produces one finished video per CSV row — 10 in Erica’s example — each with the same visual template but different question/tip text pulled from ChatGPT’s output.
- Review the batch and adjust font size on any video where text runs off-screen; Erica does this pass after generating the first batch rather than before.
Canva Pro is required for the Bulk Create app to work. Note: Bulk Create requires a Canva Pro subscription.
Do you need to appear on camera to use this method?
No. Erica pointed out that if you don’t want to be on camera, you can search Canva’s built-in video library, drag a stock clip in as the background, resize it, and send it to the back layer — the same tip text will still populate on top of it through Bulk Create. She also noted you can batch several outfit changes in one filming session (she described filming eight outfits in one day) to keep a supply of template videos ready to rotate.
Can you repurpose existing content instead of prompting ChatGPT from scratch?
Yes. Erica described uploading a PowerPoint presentation from a class she’d taught into ChatGPT and asking it to pull tips from the deck into the same question/tip CSV format, which let her promote an upcoming workshop through the finished clips. The same approach works with a blog post — feeding ChatGPT a blog and asking it to extract questions and pull-quotes in the same column structure to generate a batch of videos from existing written content.
Full video transcript
Kelly Tenny opens the hack session from the Scale with AI Summit, welcoming Erica Hannah, whom she’d interviewed previously on the Women Who AI Summit. Kelly introduces Erica as a six-time Emmy winner and a lover of people, creativity, storytelling, video, and AI. Erica explains that video and micro-content storytelling is her focus, and that video intimidates people because it seems complicated — so she wants to share an easy hack for producing YouTube Shorts, Facebook Reels, or Instagram Reels quickly.
They move into Canva, which Erica notes is affordable, well-rounded, and widely used already. She explains the method works for any niche or vertical: you go into ChatGPT and prompt it for video marketing tips in a conversational, easy-to-understand format, structured as question, tip one, tip two, tip three — then ask for 10 examples with each question and tip in its own column, delivered as a CSV file. While ChatGPT processes the request (slower than usual, Erica notes, “because of all the thinking it’s doing”), they chat about Canva’s affordability and its Magic AI features, which Erica says she’s dabbled with only a little.
After a couple of technical hiccups — ChatGPT briefly failing to load the conversation — Erica downloads the resulting CSV. She then switches to Canva and creates a new design, selecting “YouTube Short” so the canvas auto-sizes correctly. She explains that ahead of the session she filmed herself in a silent template video — waving, pointing, delivering “tip one, tip two, tip three” gestures with no audio — and that this format works well for people who dislike hearing their own voice on camera. She notes you can also use a Canva stock video as the background instead of filming yourself.
Erica adds a text heading for “Question,” matching it to the CSV column of the same name, and sets a white background effect so it reads clearly. She turns on “show element timing” to see where the question appears in her gesture video, then drags the text’s timing to align with the footage — repeating the process for “Tip 1” and “Tip 2” (her clip only included two tips, so she skips a Tip 3 layer and instead adds a closing “Good luck” text).
With the template built, Erica opens the Canva App Store and searches for “Bulk Create,” describing its icon as “boring” but the tool itself as powerful. She uploads the CSV file, then connects each text layer to its matching column: Question to Connect Data > Question, Tip 1 to Connect Data > Tip 1, and so on, skipping Tip 3 since it isn’t used in this template. Bulk Create does require a Canva Pro subscription. Once every layer is connected, she clicks Continue, then “Generate 10 designs” — and Canva produces 10 fully finished videos, each with different tip text pulled from the CSV, using the same visual template.
Erica notes you’ll likely need to tweak font sizing afterward if text runs off-screen for longer tips, and that she typically does this pass after watching the first generated video rather than trying to guess sizing beforehand. Kelly reacts with enthusiasm, noting how overwhelming video content creation usually feels — captioning, pacing, brand alignment — and how this method removes most of that friction.
Erica adds that you don’t need to be on camera at all: you can search Canva’s video library for a stock clip, drag it into the template, resize it, and send it to the back layer as the background, and the CSV-driven tips will still populate on top. She also explains that she batches filming sessions — filming multiple outfit changes in one day — to build up a stock of template videos she rotates through every few days.
She shares a further repurposing use case: she recently uploaded a PowerPoint presentation from a class she was teaching directly into ChatGPT and asked it to extract the same question/tip format from her existing slides, then used the resulting videos to promote an upcoming workshop by adding a line like “you’ll learn this and more at the workshop on Thursday.” Kelly suggests the same approach could work by feeding ChatGPT a blog post, pulling out questions, tips, or pull-quotes to create a batch of scannable videos summarizing longer written content — Erica agrees enthusiastically and says she plans to try it immediately after the session.
Kelly closes by thanking Erica for another simple, practical hack, noting that most viewers already use Canva (many with Pro accounts already), and that the method removes the usual overwhelm of video editing. Erica points viewers to YouTube tutorials searching “bulk create reels Canva” for more detail on the process. Kelly thanks Erica and wraps the session, encouraging viewers to try the hack right away.