Scaling B2B Digital Advertising Through AI-Enhanced Creative Workflows
Leveraging Figma AI and design systems to deliver 100+ ad variations in 10 days—60% faster
Overview
I began a professional relationship with Tricentis as a consultant in 2017. In the years since I have worked on most aspects of their marketing efforts including video, animation, event booth design, and all sorts of collateral design.
Tricentis needed to launch multiple digital advertising campaigns simultaneously across LinkedIn, display networks, and social platforms. Traditional design workflows would require 3-4 weeks to produce the volume of creative variations needed for effective A/B testing and multi-platform optimization.
The challenge wasn’t just quantity—it was maintaining brand consistency and high-end execution while delivering diverse concepts across 6+ platform specifications, each requiring multiple headline and CTA variations for performance testing.
Challenges
- Client required 10-12 distinct creative concepts across multiple product lines
- Each concept needed adaptation for 6+ platform specifications (LinkedIn 1200×627, Instagram 1080×1080, Display 728×90, etc.)
- A/B testing demanded multiple headline and CTA variations for each design
- Traditional workflow timeline (3-4 weeks) didn’t align with campaign launch deadlines
- Maintaining Tricentis brand standards while producing 100+ unique assets at scale
The AI-Enhanced Approach
Rather than replacing traditional design craft with AI, I integrated Figma’s AI tools strategically at specific workflow bottlenecks—accelerating ideation and scaling production while maintaining creative control and brand consistency.
Workflow Breakdown
Phase 1:
Rapid Concept Generation
Tool: Figma AI Make
Generated 10 initial layout concepts in 2 hours using targeted prompts for enterprise B2B aesthetics. Selected 3 strongest directions for brand refinement.
Time saved: 2 days → 2 hours
Phase 2:
Systematic Variation
Tool: Figma Variables + ChatGPT
Built design system with variables for headlines, CTAs, and visual elements. Generated 15 headline variations and 10 CTA options, then systematically combined them.
Output: 50+ unique combinations from 1 master design
Phase 3:
Platform Adaptation
Tool: Figma Auto Layout
Single master design automatically adapted to 8 platform specifications using Auto Layout. Maintained pixel-perfect execution across all sizes.
Result: 1 design → 80+ platform-specific assets
Before/After Workflow Comparison
Traditional Workflow
Sketch concepts: 2-3 days
Create variations: 5-7 days
Resize for platforms: 3-4 days
Export & prep: 2-3 days
Total: 3-4 weeks
AI-Enhanced Workflow
AI concept generation: 1 day
Refinement & brand alignment: 2 days
Variation scaling: 2 days
Platform adaptation: 2 days
Export & delivery: 3 days
Total: 10 days
Output
Results
100+
Unique ad
variations delivered
60%
Reduction in
production time
8
Platform specifications
from single source
100%
Brand
consistency maintained
Key Learnings
AI as Creative Accelerator, Not Replacement
Figma AI enabled rapid exploration and systematic scaling, but human creative direction remained essential for ensuring brand alignment, maintaining visual quality, and making strategic decisions about messaging hierarchy and visual impact.
Scalable Systems Over One-Off Solutions
By building a variable-based design system rather than individual assets, the workflow became repeatable for future Tricentis campaigns—enabling the client to maintain creative velocity without requiring designer involvement for every variation.
Strategic Integration Points
The key wasn’t using AI everywhere, but identifying specific workflow bottlenecks where AI provided maximum value: initial concept generation (too slow manually), copy variation (needed volume for testing), and platform resizing (tedious and error-prone).
“I’ve worked with Jordi at three highly successful B2B tech companies from $5M to $250M ARR. Whether it’s designing a brand identity, putting together product explainer videos, or tweaking UI, he’s always exceeded our product expectations.”
