EV dealership launch, Metro Vancouver
Launch Campaign & AI Ad Creative
An EV dealership launch on Meta, from catalogue migration to campaign build, plus lead-gen work and a generative AI creative workflow across the dealership group.
- Role
- Campaign strategy, creative and media buying
- Timeline
- 2026
- Tools
- Meta Ads Manager, Midjourney
- Type
- Employer

Client details are confidential. Names, logos and data are removed, and some screens are recreated or redacted.
The problem
A new EV dealership in Metro Vancouver needed a full launch on Meta, and the wider dealership group needed lead generation that worked.
My role and constraints
I come up with campaign and creative ideas, build the creatives, and build and launch campaigns in Meta Ads Manager. AI helps pull reports and flag what is underperforming; I then research and decide what to keep or cut.
Process
- Pre-launch audit and a 90-day marketing plan.
- Migrated 48 vehicles into a dedicated catalogue with dynamic product ads.
- Test-drive auto-responder emails.
- GenAI creative workflow: Midjourney plus prompt systems for on-brand ad creatives at 4:5 and 9:16.
Launch structure
Stage 1
- Pre-launch audit
Stage 2
- 90-day marketing plan
Stage 3
- Test-drive carousels
- Dynamic product ads. 48-vehicle catalogue
- Click-to-call
- Broad Advantage+
Stage 4
- Leads
Stage 5
- Auto-responder. Test-drive emails
Creative and reporting loop
Idea
I come up with the campaign and creative ideas.
Create
Midjourney plus prompt systems for on-brand creatives at 4:5 and 9:16.
Launch
Built and launched in Meta Ads Manager.
Flag
AI pulls the reports and flags what is underperforming.
Decide
I research the flags and decide what to keep or cut.
The solution
- Full launch campaign: test-drive carousels, catalogue ads, click-to-call and broad Advantage+.
- Lead-gen across the dealership group: instant forms cut from 15 to 12 fields, plus qualifier ad sets.
Outcome
- Vehicles in the launch catalogue
- 48
- Instant form fields
- 15 to 12
- Lowest cost per lead
- $26
What I learned
AI is fast at flagging what is underperforming, but the keep or cut decision still needs research and judgement.