HEMA, a Dutch retailer with a century-long history, created the internal AI assistant HAL based on Amazon Bedrock AgentCore. According to the AWS Machine Learning Blog, the system transforms information searches across multiple developer portals into instant answers.
HAL uses the Model Context Protocol and provides controlled corporate knowledge within tools that employees already use. No AWS credentials are required on client devices, and security is linked to Microsoft Entra ID.
The practical implication of this approach is to reduce hopping between work systems and embed knowledge access into familiar workflows. However, the available material is only a synopsis of an AWS publication, so details on architecture, usage scale, and measurable results remain unknown.
editorial commentary
Why it matters
The likely consequence is a further shift of corporate assistants from separate portals to embedded knowledge access within work systems. The next observable signal will be HEMA data on user numbers, connected tools, and implementation outcomes. Significant uncertainty remains because currently only an AWS synopsis is available without independent confirmation.