AWS Partner Network (APN) Blog
Accelerating supply chain AI with Kinaxis MCP on Amazon Bedrock AgentCore
By: Shriram Bidkar, Assistant Vice President, AWS Solutions – Genpact
By: Naresh Rajaram, Sr. Partner Solutions Architect – AWS
 |
| Genpact |
 |
The biggest barrier to supply chain AI isn’t the technology, it’s integration. Genpact’s Kinaxis MCP Server, built on Amazon Bedrock AgentCore, replaces months of custom API work with a single secure interface. One manufacturer achieved an estimated reduction in AI deployment time from 12 weeks to 3 days. A global retailer saw up to an 18% improvement in forecast accuracy. Discover how MCP standardization unlocks supply chain AI at scale.
A Fortune 500 consumer goods manufacturer wanted to use AI agents to support supply chain planning. Teams spent months building and securing custom API connections between the agents and their planning systems. Each new agent required additional integration work, driving up costs and delaying deployment. As a result, the engineering effort was consumed by integration tasks rather than supply chain optimization.
This scenario is the norm, not the exception. Enterprises adopting ML-based supply chain forecasts typically cut errors by 20–50%, releasing $20 to $40 million in working capital per $1 billion of revenue. The challenge isn’t acquiring AI, it’s connecting AI agents securely to the specialized systems that run modern supply chains. With industry leaders achieving up to a 30% reduction in inventory costs while improving service levels, organizations can’t afford multi-month integration projects. They need AI operational now.
Breaking down the integration barrier
Model Context Protocol (MCP) is changing this. APIs standardized how software systems communicate, now MCP is doing the same for AI agents—replacing dozens of custom integrations with a standard interface. With thousands of AI agents already built across Amazon organizations since 2025, and enterprises managing supply chains spanning hundreds of millions of SKUs, the conditions are right for a breakthrough
When AI is applied effectively in supply chain planning and execution, organizations can unlock meaningful improvements in cost, inventory efficiency, and customer service. Integration is the first step to achieving these outcomes.
How Genpact built an enterprise-ready MCP server for Kinaxis
Kinaxis RapidResponse is the system of record for demand forecasting, inventory management, and Sales and Operations Planning (S&OP) at Fortune 500 manufacturers. Despite its capabilities, connecting AI agents to RapidResponse required months of custom API development, blocking AI adoption at scale.
Working as an AWS AI Services Competency and Amazon Bedrock AgentCore launch partner, Genpact built the Kinaxis MCP Server—a production-ready integration layer that runs on AgentCore Runtime. It implements the MCP standard with OAuth 2.0 authentication, creating a single secure interface that any AI agent can use to access Kinaxis RapidResponse, collapsing timelines from months to hours.
Why Amazon Bedrock AgentCore
Genpact selected Amazon Bedrock AgentCore as the foundation because supply chain operations demand enterprise-grade reliability. AgentCore delivers secure, serverless hosting with complete session isolation—every interaction runs in its own dedicated microVM.
Without this isolation, a single compromised agent session could expose data across your entire supply chain. AgentCore helps ensure that a procurement agent accessing supplier data is completely isolated from a demand planning agent querying forecast models, even when both run simultaneously.
Figure 1 – Diagram of a supply chain AI integration showing secure agent connectivity. Users connect to MCP Server, which routes requests to Kinaxis RapidResponse using AgentCore.
Flexible deployment and seamless authentication
Organizations can deploy the Kinaxis MCP Server through AWS Marketplace using pre-built Docker images or can customize their deployment using Amazon Elastic Container Service (Amazon ECS) or Amazon Elastic Kubernetes Service (Amazon EKS) for specific compliance requirements. Amazon Bedrock AgentCore Identity integrates with Amazon Cognito, Okta, and Microsoft Entra ID. Planners authenticate once and AI agents access Kinaxis securely on their behalf, eliminating credential exposure.
Core capabilities
The Kinaxis MCP Server includes three components built to assist supply chain planning teams:
- Workbook data access: Retrieve demand forecasts, inventory positions, and production schedules directly from Kinaxis workbooks
- Targeted worksheet extraction: Access specific planning parameters without processing entire workbooks
- Advanced parameter customization: Query across site groups, time periods, and multi-dimensional planning hierarchies
All components integrate with Amazon CloudWatch for real-time visibility into agent activities and system performance.
Results
The manufacturer mentioned at the start of this post achieved an estimated reduction in AI agent deployment from 12 weeks to 3 days—a 96% reduction. A global retailer saw up to an 18% improvement in forecast accuracy, reducing stockout rates by an estimated 23% network wide. A pharmaceutical company found that OAuth 2.0 authentication and session isolation provided the security posture needed for internal approval, unblocking AI initiatives that had stalled for months.
Organizations are also discovering a compounding benefit: teams that start with demand forecasting agents expand to inventory optimization and scenario planning—all on the same integration layer. Each new use case can be deployed in days, not months, because the integration work is already done.
Getting started
If your organization runs Kinaxis RapidResponse and AWS today, you’re closer to production-grade supply chain AI than you think. The Kinaxis MCP Server is designed to significantly reduce the integration work that typically takes 3 to 6 months of engineering to days. Organizations have been able to go from zero to deployed AI agents in days, not quarters, while competitors are still scoping their integration projects.
The Kinaxis MCP Server is available now on AWS Marketplace. For a customized deployment, Genpact offers 4-week assessments to design architectures matching your specific security and operational requirements. Contact your AWS account team or Genpact representative to get started.
Genpact – AWS Partner Spotlight
Genpact (NYSE: G) is a global professional services firm with more than 125,000 people across over 30 countries. As an AWS Generative AI Competency Partner and AgentCore launch partner, Genpact stands at the forefront of enterprise AI transformation in supply chain, finance, and operations.