Accelerating your marketing ideation with generative AI – Part 2: Generate custom marketing images from historical references

2026-02-04 · Source: Artificial Intelligence · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Data Science & Analytics, Marketing, Branding & Advertising · Depth: Intermediate, long

Summary

AWS has developed an advanced image generation system that integrates generative AI with historical marketing campaign data to streamline content creation and maintain brand consistency. This solution, building on previous work with Amazon Nova foundation models, uses Amazon Bedrock, AWS Lambda, and Amazon OpenSearch Serverless. It processes historical campaign assets by enriching them with metadata and AI-generated descriptions, then transforms them into embeddings using Amazon Titan Multimodal Embeddings. These embeddings are stored in an OpenSearch Serverless index, enabling semantic search for relevant reference images based on new campaign requirements, objectives, and target audiences. The system then uses Amazon Nova Pro to generate enhanced prompts, which are fed into Amazon Nova Canvas to create new marketing visuals aligned with both user input and past successful campaigns. Bancolombia has successfully implemented this approach, reducing iterative processes and ensuring visual consistency.

Key takeaway

For marketing professionals tasked with rapidly producing engaging, brand-consistent content, this AWS solution offers a robust framework. You can leverage historical campaign data and generative AI to significantly reduce creative iteration cycles and ensure new visuals align with proven strategies. Consider adopting this architecture to scale your marketing efforts while maintaining quality and consistency, potentially freeing up resources and improving campaign performance.

Key insights

Integrating historical campaign data with generative AI enhances marketing content creation and ensures brand consistency.

Principles

• Historical data improves AI-generated content relevance.
• Semantic search aligns new content with past successes.
• Metadata enrichment boosts content generation accuracy.

Method

The method involves uploading historical assets to S3, processing them with Lambda functions to generate descriptions (Nova Pro) and embeddings (Titan Multimodal Embeddings), indexing these in OpenSearch Serverless, and then using semantic search to inform prompt engineering (Nova Pro) for new image generation (Nova Canvas).

In practice

• Use Amazon S3 for historical asset storage.
• Employ OpenSearch Serverless for vector similarity search.
• Integrate Amazon Nova Pro for prompt enhancement.

Topics

• Generative AI for Marketing
• Image Generation
• Vector Search
• Amazon Bedrock
• Prompt Engineering

Code references

• aws-samples/generative-ai-ml-latam-samples
• dlaredo/generative-ai-ml-latam-samples

Best for: Marketing Professional, AI Product Manager, Machine Learning Engineer

Related on AIssential

• How generative AI is reshaping personalized email campaigns — Generative AI enables hyper-personalized email campaigns by localizing content, scaling variations, and optimizing delivery times.
• 5 steps to generate consistent brand images with Midjourney — Consistent AI image generation relies on structured workflows, visual language, and iterative refinement.
• The 9 AI Tools Marketers Use to Create Images and Video in 2026 — The marketing industry is rapidly adopting AI tools for visual content, prioritizing commercial rights and integrated workflows.
• Amazon adds AI-generated product previews to search results — Amazon is integrating AI-generated visual previews to enhance product discovery and address imprecise user search queries.
• How Gen AI is Disrupting B2B Buying Decisions — Generative AI is fundamentally shifting B2B buying into AI-mediated "dark funnels," demanding new go-to-market strategies.
• Reimagine marketing at Volkswagen Group with generative AI — Generative AI can accelerate marketing content creation while ensuring strict brand and technical compliance through automated validation.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Artificial Intelligence.