This release delivers focused enhancements across all agents—including Account, Catalog, Inventory, Order, Price, Review, Loyalty, CSR, and Promotions—boosting efficiency, data accuracy, and workflow performance. 

Infosys Equinox Commerce AI Studio

Introducing Equinox Commerce AI Studio, a powerful, unified platform designed to transform e-commerce operations through intelligent automation. Built with modular precision and enterprise-grade flexibility, the studio empowers teams to configure AI infrastructure, deploy persona-based agents, and automate multi-step business workflows—all from one seamless interface.

Management Module: Foundation for Intelligent Automation

The Management Module provides all essential configurations required for agents and workflows to operate efficiently. It includes robust setup options for:

MCP Servers

Enable secure, controlled interactions between AI agents and internal commerce services like product, pricing, and administration.

  • MCP Servers: MCP Servers enable secure, controlled interactions between AI agents and internal commerce services like product, pricing, and admin systems. They support Header and OAuth2 authentication, allow configuration of endpoints and scopes, and manage store-level or global visibility.

LLM Providers

Integrate with a wide selection of chat, embedding, and multi-modal models via LiteLLM.

  • LLM Providers: LLM Providers integrate multiple chat, embedding, and multi‑modal models via LiteLLM, allowing configuration of model types, credentials, and descriptions.

Knowledge Bases

Enhance agent intelligence with domain-specific information repositories.

  • Knowledge Bases: Knowledge Bases boost agent intelligence by storing domain-specific information, supporting multiple repositories, chunking strategies, and embedding models. They allow single or bulk document uploads that are automatically indexed for retrieval.

Guardrails

Ensure safe, compliant AI interactions.

  • Guardrails: Guardrails ensure safe, compliant AI interactions by validating inputs and outputs using pre-call, post-call, or real‑time monitoring modes. They support configurable JSON-based rules and thresholds for content safety.

Agent Module: Build Persona-Driven AI Assistants

The Agent Module delivers the ability to create intelligent, task-focused agents that serve as chat assistants or workflow components.

Agent Creation

Configure agent personality, scope, orchestrator eligibility, and responsibilities:

  • Assign system prompts, guardrails, knowledge bases, memory models
  • Enable access to selected MCP servers
  • Choose store-level or global availability

Pre-Built Personified Agents

Jumpstart intelligent operations with ready-to-deploy agents, such as:

  • Catalogue Manager
  • Price Manager
  • Inventory Manager
  • Promotion Manager
  • Order Manager
  • Account Manager
  • Loyalty Manager
  • CSR Manager
  • Merchandize Manager
  • Subscription Manager
  • Notification Manager
  • Admin Assistant
  • Shopping Manager

These agents leverage MCP tools, provide structured responses, and execute domain-specific workflows with precision.

Workflow Module: Automate Complex Commerce Processes

The Workflow Module makes it effortless to design and execute multi-step processes without repetitive inputs.

Workflow Creation

Users simply describe a task. The system:

  • Interprets intent
  • Identifies required agents
  • Generates a workflow plan using Supervisor and Workflow agents
  • Suggests relevant resources
  • Allows user review and approval

Workflow Execution

Workflows can be saved, tested, or activated. For example, Price & Inventory Audit Workflow

  • Fetch SKUs and product information
  • Perform real-time price checks
  • Verify inventory levels
  • Generate a complete audit report

All completed automatically—no step-by-step user intervention required.

Revision History
2026-03-17 | JP – Added release notes for 8.21.5.