[Success Story] Visualizing the AX Platform Architecture:
Intuitive Understanding of Technology Achieved with 12,000 Images

1. Project Background and Pain Points

For large IT service companies like LG CNS,AX (AI Transformation) Consultingthe biggest hurdle was the 'complexity of technology'.

  • Abstract Concepts:Explaining advanced technical concepts such as hybrid clouds, AI agents, and smart city metaphors to clients (decision-makers) was limited when relying solely on text and charts.

  • Limited Design Resources:Maintaining brand guidelines throughout a consulting report exceeding 100 pages, while manually creating over 500 high-quality illustrations tailored to each page's context, was extremely costly and time-consuming.

  • Unsuitability of Existing AI Tools:General generative AI APIs had limitations in accurately visualizing a company's unique linear line art style or specialized terminology.

2. Solution & Process

We went beyond simple image generation and solved the problem by building a'Vibe Coding'-based Image Production Pipeline.to solve the problem.

  • Step 1. Reverse Engineering and Data Extraction: We used vibe coding to analyze existing websites and guidelines, converting the brand identity into digital resources.

  • Step 2. Designing a 2-Stage AI Pipeline:

    • Stage 1 (LLM): Analyzing the consulting text, it recommends 10 illustration scenes tailored to the context and generates detailed prompts.

    • Stage 2 (Imagen4): Based on the analyzed data,Google Imagen4 APIis called to mass-produce images that adhere to a fixed style guide.

  • Step 3. Style Fine-Tuning: We set up 10 fixed style combinations, such as futuristic, human-centered, and tech metaphor, to maintain a consistent visual appearance across all pages.

  • Step 4. Human-AI Collaboration (HIL): The best results from the generated images were selected, and gradients and details were corrected using Figma to complete the final quality.

3. Outcomes

  • Overwhelming Productivity: Within a short period,a total of 12,000 image candidateswere generated, securing over 500 high-quality illustrations optimized for the consulting report.

  • Improved Decision-Making Speed: By visualizing complex architectures (On-Premise, Cloud, etc.) and user intent detection processes with intuitive line art, customer understanding was maximized, leading to faster decision-making.

  • Technology Assetization: By building a company-specific image generation engine (Shot Recommendation Engine), anAX Visualization Roadmapwas secured, which can be immediately applied to other projects in the future.