Free Knowledge Case Study: Using Generative AI on AWS to Automate Insights, Cut Costs, and Scale Smarter

Introduction

Every organization uses data. It provides insights into your business but when it piles up, it often does the opposite.

With outdated data workflows, you end up collecting more reports, searching through multiple documents, and even guessing where a particular piece of information was stored.

It’s slow, it’s manual, and it gets worse the more your organization grows and the data that was meant to help you move faster ends up slowing down your processes.

This case study focuses on an implementation that transforms traditional data processing and search workflows into an intelligent, AI-powered system that delivers significant business value through improved efficiency, accuracy, and scalability.

Client’s Overview

Free Knowledge Africa is a nonprofit organization that seeks to promote free and open knowledge by Africans in travel, sports, tourism, geography, and cultural heritage towards sharing and contributing to global knowledge.

They promote the use, contribution and increase the reach of Wikimedia projects as it pertains to gaps in African travel, geographic information, sports, cultural and tourist sites.

The Challenges

The client relied on conventional data processing and search methodologies that:

• Required significant manual intervention
• Had limited scalability as data volumes grows
• Lacked intelligent query processing capabilities
• Resulted in high operational costs and slow time-to-insight

Solution Architecture by CloudPlexo

Our implementation leverages a serverless, event-driven architecture centered around AWS Generative AI services, specifically.

Core Components:

1. API Gateway: Serves as the entry point for client requests, providing secure, scalable access to the backend services

2. Primary AWS Lambda Function: Processes incoming requests and orchestrates the workflow between various components

3. Storage Layer:

   • Amazon DynamoDB: Stores structured data and metadata for quick access
   • Amazon S3: Houses large objects, documents, and media files

4. Generative AI Components:

   • Amazon Bedrock: Powers the core generative AI capabilities, including:
     • Text generation and summarization
     • Content transformation
     • Semantic understanding of queries
   • Secondary AWS Lambda Functions: Handles specialized AI processing tasks

5. Amazon OpenSearch: Provides enterprise search capabilities, enhanced by the AI processing layer for improved relevance

Data Flow and Connection

Here’s how the system works, step by step:

• Client submits requests through the API Gateway
• Primary Lambda function processes and routes the request
• Data is retrieved from DynamoDB or S3 as needed
• Bedrock applies generative AI capabilities to transform, enhance, or analyze the data
• Results are processed by additional Lambda functions and/or enriched with OpenSearch
• Final response is returned to the client via API Gateway

Results

With CloudPlexo’s implementation, Free Knowledge now achieves the following:

• The data query processing time reduced from 8.5 seconds to 1.2 seconds which is an 85% improvement.

• Data processing capacity increased from 10,000 records/hour to 250,000 records per hour, indicating a 25% boost

• Search relevance accuracy improved from 67% to 94% which is a 40% increase

• Free Knowledge’s infrastructure costs dropped from $12,500/month to $4,800/month. That’s 62% cost reduction

• Manual intervention reduced from 450/month to 25/month.This is 94% drop in manual work

• Time-to-Market for New Features:
  Accelerated from 12–16 weeks to 2–3 weeks — an 80% reduction in rollout time.

AWS Generative AI Services Used

CloudPlexo leveraged the following AWS Generative AI-specific services in implementing this project:

Amazon Bedrock

• Implements foundation models to power natural language understanding
• Provides contextual comprehension of industry-specific terminology
• Enables semantic search capabilities through vector embeddings
• Facilitates document summarization and information extraction

AWS Lambda with AI/ML Extensions

• Custom processing of Bedrock outputs for industry-specific applications
• Real-time transformation of input data for AI processing
• Intelligent routing of requests based on content analysis

Amazon OpenSearch with Vector Search

• Stores and retrieves vector embeddings generated by Bedrock models
• Enables semantic search capabilities beyond traditional keyword matching
• Provides relevance ranking based on AI-powered understanding

The new system implemented by CloudPlexo replaced time-consuming manual tasks with smart automation.

So, instead of relying on people to sort documents, search by exact keywords, or follow rigid data rules, the upgraded process now understands content, finds relevant answers based on intent, and spots patterns on its own which makes everything faster, more accurate, and much easier to manage.

Also, the technical setup was completely modernized. The old, bulky system was replaced with a flexible, serverless structure that can scale up or down as needed.

Maintenance became easier with fully managed services, and the development process got smarter too, using AI-driven tools that continuously learn, simplify code, and improve performance over time.

Conclusion

This transformation is a complete shift in how data is processed, searched, and used. By replacing manual, slow processes with intelligent automation, the system now works faster, smarter, and more efficiently than ever before.

If you’d like to implement similar results for your organization, CloudPlexo can help you build scalable, intelligent solutions that drive real impact. Visit www.cloudplexo.com to get started.