For most tasks, yes. Approving composition first avoids expensive video re-generations. The exception is quick B-roll clips where you already have saved Characters or Products — those can go direct to video.
How-to guide
Choose the Right Approach
Pick the right path in under 10 seconds and reduce wasted credits.
⏱ 3 min readTL;DR: Match task type to workflow: still image, video, audio/talking, strategy, adaptation. Start in Chat, move to Studio for precision/batch.
Decision by Task Type
| Task Type | Best Path | Notes |
|---|---|---|
| Still image | Chat-first | Fast conversational iteration |
| Video | Keyframe first → Animate | Cheaper and safer than full-video retries |
| Audio / talking content | Portrait → Script → Voice → Talking video | Best for spokesperson or UGC-style flow |
| Strategy / planning | Direct planning prompts in Chat | Calendar, campaign plan, copy variants, A/B plans |
| Adaptation / final delivery | Enhance → Resize | Finish quality first, then platform ratios |
Chat vs Studio
| Use Chat when… | Use Studio when… |
|---|---|
| You want strategic guidance and iterative conversation | You need direct parameter control |
| You're exploring direction | You're running repeatable/batch production |
| You need end-to-end flows | You need maximum precision |
Default: most users should start in Chat.
Quality Tiers
- Fast: drafting, concept tests, social iterations.
- High quality: final deliverables and campaign assets.
- Cinematic/premium: hero pieces and brand films.
Rule: Start fast, upgrade once direction is approved.
Copy-Paste Decision Prompts
Workflow Decision
I want to create [goal]. What is the best path: still-first, direct video, or another workflow? Platform: [platform]. Need: [draft/final]. Return: recommended path + first prompt + fallback option.
Step-by-Step Guide
I need [deliverable] for [platform]. Subject: [subject]. Style: [style]. Ratio: [ratio or recommend]. Need consistency: [yes/no]. Return a step-by-step workflow and prompts.
Multi-Platform Adaptation
Adapt this asset for all channels: 1:1 (feed), 9:16 (stories/tiktok), 16:9 (youtube/linkedin). Keep subject centered and text within safe zones.
Quick Reference Card
| Need | Fast command pattern |
|---|---|
| Still image | "Create a [scene] for [platform]." |
| Video | "Generate keyframe first, then animate." |
| Resize | "Resize this to 9:16 / 1:1 / 16:9." |
| Enhance | "Enhance this for final quality." |
| Voice/speech | "Create a voiceover for this script: [text]." |
| Talking head | "Portrait + script + voice → combine." |
| Campaign plan | "Plan a campaign for [goal]." |
| Content calendar | "Create a [timeframe] calendar for [brand]." |
Save the person as a Cast member in your Elements library with reference photos and a description. Then reference them by name in future prompts — Madison will maintain consistency automatically.
No. The Resizer uses AI-powered canvas extension rather than simple stretching. It generates new context around your image to fill the new aspect ratio while keeping the focal point intact.
Typically, individual clips are 5–8 seconds. For longer videos, build your sequence by chaining multiple scenes together. This gives you more control over pacing and transitions.
Most mismatches come from prompt quality. Check that you've specified aspect ratio, subject definition (Product vs Cast), and creative direction. Try changing one variable at a time rather than rewriting the entire prompt.
Products are physical items (shoes, bottles, packaging). Cast are people or characters (founders, models, mascots). Both are managed in Elements for consistency, but they're treated differently in composition and scene generation.
What to try next
Ready to create?
Start generating professional ads with Madison in minutes.
Free to start. No credit card required.