​
• Ekyam Platform Overview

Ekyam Platform is built upon a proprietary data platform engineered to transform  raw data into an AI-ready format. At its core, it provides a comprehensive library of AI agents, each specialized in retail operations, designed for seamless collaboration amongst themselves and with other 3P AI agents. Ekyam Platform enables real-time communication and data synchronization across all connected retail systems, regardless of their underlying technology or data structure, offering a solution with its AI-powered retail integration platform designed to make your data work smarter.

​
• Platform Architecture

Group2085666857 Pn

​
Ekyam Universal Connector

It is a versatile Ekyam component, built as microservices, that supports all common protocols and formats (APIs, FTP/SFTP, databases, XML/JSON files, EDI X12/EDIFACT, etc.) with no-code configuration. Each connector ingests or sends data in any format and automatically to Ekyam’s canonical schema. The external systems (ERP, OMS, WMS, POS, ecommerce, etc.) are connected via Ekyam’s pre-built connectors, which extract data (orders, shipments, inventory updates, vendor info, etc.) in real time. These connectors feed data into the core platform. Ekyam’s integration layer “serves as a bridge” between disparate retail systems, providing a unified, real-time view of inventory and orders.
Key Features
  • The connector layer is extensible – custom connectors and mappings can be added – and fully monitored (tracking throughput, errors, SLAs). For instance, connecting an ERP’s flat-file export takes minutes by configuring endpoints; the system then uses AI-driven profiling to map fields and validate formats.
  • The connector also handles bi-directional flows: outbound documents are generated from the Chronicle data and sent to partners.  All flows are logged and audited to ensure visibility. 
  • With built-in monitoring, throttling, retry logic, and anomaly detection, the Universal Connector ensures data integrity, availability, and security across the entire retail stack.
Security Highlights:
  • End-to-end encryption for all data in transit and at rest (TLS/SSL, AES-256)
  • Token-based and certificate-based authentication
  • Role-based access control (RBAC) and data-level permissions
  • Audit trails and activity logging for every transaction and event
  • Compliance-ready architecture aligned with SOC 2, GDPR, and ISO 27001 best practices.

​
Ekyam Universal Reader

An Ekyam component responsible for ingesting raw data from various systems, including parsing EDI/iDOCs and transforming it into Ekyam Data Standards.
Key Features
  • It has the intelligence to process an old EDI file, a new API feed or a PDF invoice. 
  • This component ensures that Ekyam’s AI-native platform can seamlessly integrate with and understand the entire spectrum of a retailer’s data ecosystem. 
  • It has the capability of converting all incoming data into an understandable format that can be used to build the Retail Knowledge Graph (RKG) and AI agents. This processed data is then available for RKG to build connections and for AI agents to make intelligent decisions and automate workflows.

​
Ekyam Universal Writer

It supports CSV, XML, JSON, TXT, etc., for internal systems, reports, or custom integrations.
Key Features
  • The “Universal Writer” ensures that data is sent automatically based on business rules or AI-driven triggers (e.g., automatically sending a shipping notification once an order is fulfilled or updating stock levels on all sales channels simultaneously). 
  • It ensures that the valuable insights and automated actions generated by Ekyam’s AI and Retail Knowledge Graph are not confined to the platform but can flow seamlessly throughout the entire retail ecosystem, enabling true operational efficiency and real-time decision-making.

​
Ekyam Universal Ledger

A centralized, real-time, standardized data store within Ekyam that acts as the definitive Source of Truth for key operational data like inventory, orders, and customer information.
Key Features
  • The Ekyam Universal Ledger is the platform’s core data backbone, creating a unified, real-time record of all critical retail operations. 
  • It consolidates retail data into a single, intelligent, and continuously updated record, empowering the entire Ekyam platform to deliver true AI-driven efficiency and visibility for retailers.

​
Retail Knowledge Graph (RKG) Core

Ekyam’s Retail Knowledge Graph (RKG) organizes all core retail data, products, SKUs, categories, vendors, inventory, orders and store locations into a unified, intelligent model. It is a graph database that models retail entities (products, SKUs, categories, vendors, locations, customers, etc.) and their relationships. The graph encodes the semantic schema (ontology) of the retail domain (for example, “Product A is supplied by Vendor X” or “SKU123 is stocked at Warehouse Y”). This structured representation provides a “system of truth” for facts about products, vendors, and static reference data. Natural language queries about product details or vendor relationships are grounded in the RKG, which can be queried via graph query APIs or natural-language-to-graph translations.
Key Features
  • The RKG will make it easier to answer business-critical queries like: “Which products in size XL sold at Store 123 last month?” or “Which vendors frequently short-ship a specific SKU?”

​
AI-Native Middleware

The top layer consists of intelligent agents and AI-driven applications. 
Key Features
  • These agents can converse, analyse, and act on the data. For instance, an agent might monitor sales velocity and auto-generate reorders or interact with a merchant in natural language to explain an inventory discrepancy. 
  • By combining RKG and vector retrieval (the “GraphRAG” approach), agents provide answers that are accurate and explainable in the retailer’s context.  They can then trigger workflows or notifications as needed. 
  • This AI-native design means Ekyam is built from the ground up to support generative AI agents that perform end-to-end retail tasks.

​
Chronicle Unified Schema

Chronicle is the powerhouse behind Ekyam, a sophisticated data engine designed to capture and make sense of all product and inventory activity across your entire retail ecosystem. Chronicle effortlessly normalizes events from diverse sources into a common format. This transforms fragmented data into something truly valuable—it becomes searchable, comparable, and incredibly insightful. This comprehensive event timeline acts as a real-time, chronological record. It empowers your AI assistant to provide instant, up-to-date answers to critical questions like “What orders shipped yesterday?” or “What returns occurred last week?” More than just answers, Chronicle ensures your assistant maintains seamless conversation context, leading to more accurate and helpful interactions.
Key Features
  • It combines structured relational storage—for facts like SKU-level stock counts, product attributes, and store inventory positions—with a vector database designed for unstructured data such as product descriptions, packaging notes, or supplier-uploaded spec sheets. 
  • All data normalized by the Retail Knowledge Graph (RKG) is written into Chronicle. For AI use cases, textual content (e.g., “color discrepancy in recent shipment” or “damage reason noted by receiver”) is broken into chunks and vectorized.
  • Vector embeddings allow Ekyam agents to semantically search past records and retrieve similar inventory events—such as SKUs flagged for quality issues or suppliers with recurring item mislabels. 
  • In Retrieval-Augmented Generation (RAG) workflows, an agent may ask, “Have we had this issue before with SKU-123 in the past six months?” Chronicle then returns semantically relevant examples, which the AI model uses to generate a grounded recommendation. Acting as the AI’s long-term memory for retail product data, Chronicle enables intelligent, explainable decision-making that links structured inventory records with real-world context.

​
GPT Style AI- Agents 

In the Ekyam platform, GPT-style AI agents are intelligent, autonomous entities that use LLM capabilities to understand complex retail data, reason over the Retail Knowledge Graph (RKG), generate human-readable insights, and act across integrated systems. They enhance automation, provide deeper insights, reduce manual effort, and improve response times. The agent has short-term context and long-term memory, as well as tool-using capability. Ekyam’s AI- Agents understands the user’s requirements, breaks it down into simple structure, creates and runs a query, analyse the data and respond to the user.

​
Vector Database (Semantic Memory)

A vector (embedding) store holds representations of unstructured or semi-structured information for long-term “semantic memory.” This includes embedding past conversations with the assistant, customer or product documents, and even summarized contents of Chronicle events. At runtime, the agent uses semantic search over the vector DB to recall relevant information (for example, pulling up similar past questions or related context). This enables the assistant to remember details across conversations and find relevant facts by meaning not just keywords.

​
Internal Tooling / APIs

A suite of internal microservices or API endpoints (called “tools” by the AI agent) for data access and mutation. Examples include GetInventoryBySKU, GetOrderStatus, SummarizeReturns, CreateRestockOrder, etc. Each tool encapsulates a backend operation (often implemented via Ekyam APIs or direct queries to the RKG/Chronicle) and exposes a defined input/output schema. The agent can invoke these tools programmatically (through the LLM’s function-calling ability or agent toolkit) to retrieve live data or trigger processes. For instance, calling GetInventoryBySKU(“SKU123”) might query the RKG or inventory database and return current on-hand quantities.