What challenges did AutoScout24 face before AI implementation?

AutoScout24 AI Platform engineers spent up to 30% of their time on repetitive support tasks such as answering how-to questions, granting access to tools, and locating documentation. The team needed a secure, scalable, maintainable way to move from fragmented experiments to a standardized enterprise AI framework.

What was the impact of the AI solution?

A production-ready Slack support bot is reducing manual support load on the AI Platform Engineering team. The Bot Factory blueprint gives AutoScout24 a reusable template to accelerate future agent development and extend the platform with new specialized agents. The approach turns fragmented experiments into a standardized framework that helps more teams build secure and scalable AI agents.

Source

Evidence: Low35/100

AutoScout24: Bot Factory to standardize AI agent development on Amazon Bedrock AgentCore

Use case typeDeveloper productivityUpdated Jan 14, 2026

AutoScout24 built a reusable Bot Factory blueprint to standardize AI agent development for internal developer support. The architecture uses Amazon Bedrock AgentCore, Amazon Bedrock Knowledge Bases, Amazon API Gateway, AWS Lambda, Amazon SQS FIFO, AWS Secrets Manager, AWS X-Ray, and the Strands Agents SDK to run secure, serverless Slack-based agents that answer documentation questions and execute tool actions like granting GitHub Copilot licenses.

Organization: AutoScout24
Industry: Automotive
Location: Germany
Published: January 2026

Planned next steps

The source says the organization aims to achieve: Standardized enterprise AI development framework.

Why do we believe this?Outcome claims, sources, and evidence checks

Normalized claim

Engineers' time spent on repetitive support tasks: 30%

AWS Machine Learning BlogJan 14, 2026Blog postExplicit claimLow evidence strength

AutoScout24 AI Platform engineers were spending up to 30% of their time on repetitive tasks

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

Customer: AutoScout24
Provider: AWS
Maturity: Unknown
Linked source: AWS Machine Learning Blog

No explicit deployment-stage evidence found.

Customer identity supportedSource describes one deploymentMaturity evidence evaluated

Primary read

Use case focus

Showing 1 of 1

1Developer productivity

They built an event-driven Bot Factory with Slack ingress, Amazon API Gateway, AWS Lambda, Amazon SQS FIFO, Amazon Bedrock AgentCore Runtime, an orchestrator agent, and specialized sub-agents for retrieval and tool use.
Amazon Bedrock Knowledge Bases provides retrieval over internal documentation, Amazon Bedrock AgentCore provides secure serverless execution and session isolation, and AWS Secrets Manager stores tool credentials for runtime access.
The orchestration pattern delegates requests from a central orchestrator to specialized agents, such as a Knowledge Base agent for questions and a GitHub agent for actions.

Technologies

Amazon Bedrock AgentCore Amazon Bedrock Knowledge Bases Amazon API Gateway AWS Lambda Amazon SQS FIFO AWS Secrets Manager AWS X-Ray Strands Agents SDK

A production-ready Slack support bot is reducing manual support load on the AI Platform Engineering team.
The Bot Factory blueprint gives AutoScout24 a reusable template to accelerate future agent development and extend the platform with new specialized agents.
The approach turns fragmented experiments into a standardized framework that helps more teams build secure and scalable AI agents.

Architecture

Slack messages flow through Amazon API Gateway to AWS Lambda for signature verification, then into an Amazon SQS FIFO queue. A second Lambda trigger starts the agent in Amazon Bedrock AgentCore Runtime. An orchestrator agent built with the Strands Agents SDK delegates to specialized agents, including a RAG knowledge-base agent and a GitHub action agent. AgentCore provides isolated per-session execution, while AWS Secrets Manager stores tokens and AWS X-Ray supports tracing and observability.

Sources & evidence1

Evidence: Low35/100Evidence strength