Normalized claim
Accuracy: 90% increase
The chatbot provides personalized, accurate answers with more than 90% accuracy, surpassing the business target of 80%.
Launching a High-Accuracy Chatbot Using Generative AI Solutions on AWS with Megamedia
Use case typeConversational supportUpdated Jun 13, 2026
Megamedia, a leading media holding company in Chile, aimed to simplify access to public government support program information and improve user experience for visitors of its news website. The company built a generative AI-powered chatbot to enable personalized queries about government benefits programs, replacing slow manual searches through multiple websites. Megamedia used Amazon Bedrock, accessed via the Generative AI Application Builder on AWS, and integrated intelligent enterprise search with Amazon Kendra to create and deploy the chatbot. The chatbot leverages the Claude 2 large language model (LLM) from Anthropic, hosted by Amazon Bedrock, to provide conversational and accurate responses. Partner ARKHO contributed AI expertise and helped refine and deploy the model on the website. The resulting Dato Útil chatbot achieves over 90% accuracy in answering user queries and reduced information search time from hours to about 30 seconds. It supports journalists to find source government information more efficiently, easing internal adoption. This solution launched publicly in May 2024 after a rapid 4-month development period.
- Organization
- Megamedia
- Industry
- Public Sector
- Location
- Chile
- Published
- April 2026
Reported outcomes
30 seconds
timeTime & speed
+90%accuracy+80%accuracy
Strategic outcomes
Customer experience & trustSimplified access to public benefits informationNew product / capabilityLaunched personalized conversational chatbotCustomer experience & trustDelivered accurate personalized answersSpeed & agilityEnabled rapid AI deployment
Why do we believe this?Outcome claims, sources, and evidence checks
Normalized claim
Accuracy: 80% increase
The chatbot provides personalized, accurate answers with more than 90% accuracy, surpassing the business target of 80%.
Normalized claim
Time: 30 seconds decrease
User search times reduced drastically from hours to approximately 30 seconds, enhancing accessibility for the public.
Why do we believe this deployment?Customer identity, provider attribution, maturity, and source checks
- Customer
- Megamedia
- Provider
- AWS
- Maturity
- Unknown
- Linked source
- AWS Customer Stories
No explicit deployment-stage evidence found.
Customer identity supportedSource describes one deploymentMaturity evidence evaluated
Primary read
Use case focus
Showing 3 of 3
- 1Generative AI Chatbot
- 2Intelligent Search
- 3Customer Engagement
- Megamedia needed to improve user experience by simplifying access to often complex and fragmented public benefit information available through multiple government sites.
- The broadcasters wanted to provide personalized, conversational answers tailored to individual user situations rather than static information pages.
- Megamedia leveraged AWS Generative AI Application Builder along with Amazon Bedrock to access large language models from Anthropic.
- ARKHO partnered with Megamedia to develop and tune the chatbot, including using Amazon Kendra to extract and embed relevant government benefits data for vector search.
- Claude 2 LLM was integrated via Amazon Bedrock for natural language understanding and conversational AI capabilities.
- The solution was embedded on Megamedia's news website to deliver an interactive chatbot experience accessible on multiple devices.
- The chatbot provides personalized, accurate answers with more than 90% accuracy, surpassing the business target of 80%.
- User search times reduced drastically from hours to approximately 30 seconds, enhancing accessibility for the public.
- Journalists benefit from easier access to verified source data, supporting faster content creation and increasing internal acceptance of AI.
- The project demonstrated that rapid AI innovation and deployment is achievable within four months, boosting organizational motivation for future initiatives.
Architecture
The architecture uses Amazon Bedrock to access Anthropic's Claude 2 LLM via the Generative AI Application Builder on AWS. Amazon Kendra is used to index and embed government benefit documents for accurate information retrieval through vector searches integrated with the chatbot interface.
Implementation partners1
Sources & evidence1
Evidence: Medium50/100Evidence strength
- Customer explicitly identified
- Primary source available
- Quantified outcome available
- Technical implementation details available
Type: Customer StoryPublished: Apr 29, 2026Publisher: AWS Customer StoriesEvidence: PrimaryConfidence: High
AI-generated summary. Verify important details with the linked sources before relying on this case.
Explore related AI use cases
Customer evidence
Provider evidence
AI patterns
Browse the catalog
Was this useful?
Community
Comments
Loading comments...