Mobileye deployed an AI agent on Amazon Bedrock AgentCore to automate internal support
Mobileye deployed an AI agent for internal support on Amazon Bedrock AgentCore that, according to the company, reduced response time by 90 % and achieved accuracy above 95 %, eliminating 66 % of routine manual queries about the processing status of driving data.
Mobileye, a manufacturer of autonomous driving systems, deployed an AI agent to automate internal support on the Amazon Bedrock AgentCore platform. The reason was the workload on the internal support team: according to the company, 66 % of tickets were routine queries about the processing status of driving data, and handling them required manually working through up to 15 steps across several internal systems – identifying sessions, cross-checking in visualization tools, verifying outputs and reviewing logs.
The agent uses models from Anthropic (Claude), accessed through the internal LLM Gateway at Mobileye, with quota management on top of Amazon Bedrock. According to the company, a key element is Model Context Protocol (MCP), which gives the agent real-time access to the API of the driving data processing platform – it can check the status of a specific session and retrieve processing logs and diagnostic information. According to the description, the agent therefore not only classifies tickets but also looks up information itself: it confirms completed sessions, including access credentials; for errors, it identifies the specific problem with a debugging recommendation and a link to the log; or it guides the user through the process of submitting a missing request – all without human intervention.
According to the company, the production deployment has a hybrid architecture: an on-premises component (Local Orchestrator) serves the internal ticketing system, which is inaccessible from AWS and was the main reason for this solution, while the agent itself runs on the serverless AgentCore Runtime in AWS with automatic scaling and a single API call for integration. According to the company, AgentCore Observability makes it possible to track the entire agent interaction, including MCP tool calls, which sped up debugging.
The company states that the proof of concept targeted classification accuracy of 95 % and responses within two minutes, and that the production deployment met or exceeded these targets – response time fell by 90 % and accuracy exceeded 95 %. The architecture description in the available text is incomplete. Details can be found in the source article.
Why it matters
The case shows a specific use of agentic AI beyond customer chat: automating internal operational queries by connecting a model to existing on-premises systems via Model Context Protocol. For companies with a similar volume of routine internal requests (task status, data processing, diagnostics), this is a documented example where the benefits can be measured in response time and freed-up specialist capacity, without having to run their own infrastructure for AI agents.
Two audiences, two different impacts
What this means
For individuals
For developers and engineers working with internal support or agent systems, this is a documented example of how to connect an AI agent to existing on-premises tools via Model Context Protocol, without it being merely a chatbot with fixed scenarios.
For a business
The company describes a specific reduction in internal support costs: automating 66 % of routine queries frees up experienced engineers for more complex tasks and reduces response time by 90 % with accuracy above 95 %, without the company having to manage its own infrastructure for AI agents.
ProductivityCheck the original
Event sources
only one source so far · 1 publisher, 0 independent. We count feeds from the same owner only once.