What challenges did Thorn face before AI implementation?

Thorn needed a more efficient way to handle internal customer support requests across IT, security, and engineering teams. The organization wanted to reduce repetitive questions and improve response times while keeping staff focused on the mission.

What was the impact of the AI solution?

The Toolsmiths team expected to save themselves multiple interruptions each week. The solution improved response time for internal users and increased productivity. Thorn said the effort helped the team focus more effectively on its mission.

Source

Evidence: Low25/100

Thorn Elevates Internal Customer Support Using Amazon Bedrock Generative AI

Use case typeCustomer support automationUpdated Oct 9, 2024

Thorn used AWS generative AI to improve internal customer support for IT, security, and engineering teams. The team built a Slack chatbot with retrieval-augmented generation to answer repetitive technical questions, summarize long discussions, and help staff become familiar with generative AI. The solution was integrated with Thorn's existing AWS environment and productionized after an initial hackathon prototype.

Organization: Thorn
Industry: Other
Location: United States
Published: October 2024

Reported outcomes

Strategic outcomes

Other strategic outcomeReduced internal support interruptionsSpeed & agilityRapidly prototyped a support chatbotOther strategic outcomeFreed staff to focus on the nonprofit mission

Why do we believe this deployment?Customer identity, provider attribution, maturity, and source checks

Customer: Thorn
Provider: AWS
Maturity: Unknown
Linked source: AWS Public Sector Blog

No explicit deployment-stage evidence found.

Customer identity supportedSource describes one deploymentMaturity evidence evaluated

Primary read

Use case focus

Showing 2 of 2

1Customer support automation
2Knowledge management

Thorn needed a more efficient way to handle internal customer support requests across IT, security, and engineering teams.
The organization wanted to reduce repetitive questions and improve response times while keeping staff focused on the mission.

Thorn's Toolsmiths team built a Slack chatbot using a retrieval-augmented generation pattern.
The solution used Amazon Bedrock with Anthropic Claude models and Amazon Kendra knowledge base context, supported by Amazon S3, Amazon EKS, Amazon OpenSearch Service, and AWS Identity and Access Management.
Thorn later added a second Slack integration for summarizing long chat discussions and created an internal web application so staff could gain familiarity with generative AI.

Technologies

Amazon Bedrock Anthropic Claude 3 Amazon S3 Amazon EKS Amazon OpenSearch Service AWS Identity and Access Management Amazon Kendra Amazon OpenSearch Serverless

The Toolsmiths team expected to save themselves multiple interruptions each week.
The solution improved response time for internal users and increased productivity.
Thorn said the effort helped the team focus more effectively on its mission.

Architecture

A Slack chatbot architecture routed user requests through a retrieval-augmented generation workflow. The implementation used Amazon S3 and Amazon EKS, Amazon Bedrock with Anthropic Claude for responses, and Amazon Kendra as the knowledge base context. Thorn later modernized the solution with Amazon OpenSearch Serverless and added a second Slack integration plus an internal web app.

Sources & evidence1

Evidence: Low25/100Evidence strength