Amazon gains the ability to process data from multiple sources with multimodal intelligence.

Amazon introduces a multi-agent engineering assistant that uses multimodal intelligence. The technology integrates multimodal embeddings and 22 specialized agents to automate complex decisions. The solution is aimed at companies that need automation in areas such as plant breeding and data security.

Amazon introduces a multi-agent engineering assistant that uses Amazon Bedrock AgentCore, Nova Multimodal Embeddings and Strands Agents SDK. Nova Multimodal Embeddings technology enables visual and textual analysis using multimodal embeddings, bringing together data in different formats. Bayer Crop Science developed an Agentic Decision Engine (ADE) based on Amazon Bedrock AgentCore to automate data analysis in plant breeding.

The ADE includes 22 specialized agents, such as the Troubleshooting Agent, which analyzes logs in CloudWatch to identify failures in workflows. Octave Relance created a database for AI on Amazon Bedrock, allowing regulated companies to search for information in documents without risk of leakage. AWS presents a blueprint for an enterprise agentic AI platform based on Amazon Bedrock and AgentCore, focused on governance, security and operations across multiple teams.

AWS published a post about building a semantic layer for agentic AI using Strategy Mosaic and Amazon Bedrock AgentCore. Mosaic's semantic layer supports multiple protocols, including MCP, JDBC/ODBC, XMLA and REST, and integrates sources such as Amazon Redshift, Aurora, PostgreSQL, Salesforce and others. The technology is developed and applied within the AWS ecosystem, with a focus on enterprise and security solutions.

For companies that need to automate complex decisions, this capability reduces the need for manual data conversion, speeding up decision-making. These agents work together to deliver a more complete and efficient response. This means that companies that use AWS can integrate the solution without difficulty, but they need to make sure their data is aligned with compatible sources, such as Amazon Redshift, Aurora, PostgreSQL and Salesforce.

Amazon redefines agent engineering with multimodal intelligence

The platform allows agents to process visual and textual data at the same time, creating a new approach to automating complex decisions. This evolution comes after years of AI agent development, in which the main difficulty was the secure and efficient integration of multiple data sources. The platform can reduce agent deployment time, allowing teams to focus on real problems rather than technical bureaucracy.

Agent engineering gains the ability to process multimodal data and automate complex decisions, redefining how systems interact with heterogeneous information. Nova Multimodal Embeddings allows the platform to understand text and images simultaneously, expanding analytical capabilities in corporate environments. Octave Reliance switched its AI database to a solution with security for regulated sectors.

Text written with AI from the sources listed and checked by the newsroom's automated rules. How we work

• #hardware
• #software
• #ia