# Data and Flow Diagrams [Take me back to main page](https://servicenow-lf.gitbook.io/the-workflow-data-fabric-loom/readme) ## Components Let us first start by breaking down the different components of the lab. The diagram below is a good representation of a tightly integrated ServiceNow landscape that spans various internal ServiceNow components and external data sources. These internal and external components will be used by Flows and AI Agents to provide the automations needed to solve our business problem of managing financial budgets. While the components will look overwhelming, the reality is customer landscapes require this level of complexity to manage different types of data across multiple functions. The key thing to note is the end user will interact with **Workspace**, **Employee Center**, or an **MCP Client** (e.g., Claude Code or Desktop). ### External system prerequisites Baseline integration to the external systems listed in this lab have mostly been configured. {% @mermaid/diagram content="graph LR subgraph "External System Prerequisites" ERP\[(ERP System
Data)] ExpenseAPI\[Expense Event
API] SharePoint\[SharePoint
Executive Memos] CDW\[(Cloud Data
Warehouse)] end ``` %% Styling classDef external fill:#2196F3,stroke:#1565C0,stroke-width:2px,color:#fff class ERP,ExpenseAPI,SharePoint,CDW external" %} ``` > [📊 View High-Resolution Diagram](https://raw.githubusercontent.com/leojacinto/WorkflowDataFabric-TypeA/main/.gitbook/assets/dataflow_prerequisites.png) * **ERP**: This lab will use an SAP system with either BAPI/RFC or OData endpoints. The authentication and integration is already configured in this exercise and the objective is get the needed data by selecting the ERP data model and extraction table. If you wish to learn more on how to create the configuration in your own environment, check this [Zero Copy Connector for ERP guide by Leo Francia in the ServiceNow community](https://www.servicenow.com/community/app-engine-for-erp-blogs/part-1-of-4-intelligent-erp-workflows-get-sap-data-into/ba-p/3192800). You can also take this ServiceNow University course on [Introduction to Zero Copy Connector for ERP Data Products and Process Extensions](https://learning.servicenow.com/lxp/en/app-engine/introduction-to-zero-copy-connector-for-erp-data-products-and?id=learning_course_prev\&course_id=72e3387d937bea54fb94b4886cba1095). * **Cloud Data Warehouse**: Snowflake will be the cloud data warehouse used in this lab. If you have a Databricks or Redshift environment, the principles and steps here will also apply. The Snowflake key-pair authentication and integration is already configured in this exercise and the objective is get the needed data asset by selecting it from Workflow Data Fabric Hub. If you wish to learn more on how to create the configuration in your own environment, check this ServiceNow University course on [Zero Copy Connector Basics](https://learning.servicenow.com/lxp/en/automation-engine/zero-copy-connector-basics?id=learning_course_prev\&course_id=c505959493283e903cc0322d6cba1025). * **Document Storage**: SharePoint will be used as the document storage which will be the target of External Content Connectors or XCC. Unstructured data will be stored in SharePoint which will be indexed by ServiceNow as additional source of data for Flows and AI Agents in this lab exercise. If you wish to learn more on how to create the configuration in your own environment, check this ServiceNow University course on [Introduction to AI Search and External Content Connectors](https://learning.servicenow.com/lxp/en/now-platform/introduction-to-ai-search-and-external-content-connectors?id=learning_course_prev\&course_id=62283c7c93d46e50f2d9bc686cba107b). * **Expense Event API**: This component can be created as a mock endpoint using services such as [beeceptor.com](https://beeceptor.com/). The specification for the API will be provided in this lab so you can simulate it in your own environment. ### User interaction layer The end user will interact with **Workspace**, **Employee Center**, or an **MCP Client** (e.g., Claude Code or Desktop). These three interfaces will be end test scenario for the lab exercises. As of March-2026 version of this lab, MCP Server access through Claude or similar is not yet included. {% @mermaid/diagram content="graph LR subgraph "User Interaction Layer" Employee((Employee/
Finance Manager)) EC\[Employee Center or
Workspace with Now Assist] ClaudeDesktop\[Claude Desktop
+ MCP] end ``` %% Styling classDef user fill:#F5F5F5,stroke:#616161,stroke-width:3px,color:#1a1a1a class Employee,EC,ClaudeDesktop user" %} ``` > [📊 View High-Resolution Diagram](https://raw.githubusercontent.com/leojacinto/WorkflowDataFabric-TypeA/main/.gitbook/assets/dataflow_user_interaction.png) ### ServiceNow Workflow Data Fabric The next diagram shows the various ServiceNow components that interact with the external systems while working in the back-end to provide the data and automation needed by users. {% @mermaid/diagram content="graph subgraph "ServiceNow AI Platform" spacer\[ ]:::hidden subgraph "Data Integration Layer" ZeroCopySQL\[Zero Copy SQL
Connection] ZeroCopyERP\[Zero Copy ERP
Connection] IntHub\[Integration Hub
Spoke/Flow] ExtContent\[External Content
Connector] MCP\[MCP] end ``` subgraph "Zero Copy Tables - Read Only" ZCCC[(Cost Center)] ZCCH[(Cost Center
History)] ZCCS[(Cost Center
Summary)] ZCExp[(Expenses)] end subgraph "ServiceNow Native Tables" ExpenseTable[(Expense Event
Line Items
Scoped Table)] FinCase[(Finance Case
Table)] end subgraph "AI & Automation" Agent1[Agent: Over-Budget
Case Creator
Zero Copy Source] Agent2[Agent: Proactive
Budget Alert
Integration Hub Source] RAG[RAG - Retrieval
Augmented Generation] FlowAction[Flow and Action] end subgraph "AI Experiences" Lens["ServiceNow
Lens"] DocIntel["Document
Intelligence"] end end %% Styling classDef external fill:#2196F3,stroke:#1565C0,stroke-width:2px,color:#fff classDef user fill:#F5F5F5,stroke:#616161,stroke-width:3px,color:#1a1a1a classDef nowassist fill:#FFB300,stroke:#F57F17,stroke-width:2px,color:#1a1a1a classDef platform fill:#4CAF50,stroke:#2E7D32,stroke-width:2px,color:#fff classDef wdf fill:#9C27B0,stroke:#6A1B9A,stroke-width:2px,color:#fff classDef hidden fill:none,stroke:none,color:transparent,font-size:0 class ZeroCopySQL,ZeroCopyERP,IntHub,ExtContent wdf class ZCCC,ZCCH,ZCCS,ZCExp wdf class ExpenseTable,FinCase platform class FlowAction platform class Agent1,Agent2,RAG nowassist class MCP wdf class Lens,DocIntel nowassist" %} ``` > **Color Legend:** 🟡 Now Assist | 🟢 Platform | 🟣 Workflow Data Fabric > [📊 View High-Resolution Diagram](https://raw.githubusercontent.com/leojacinto/WorkflowDataFabric-TypeA/main/.gitbook/assets/dataflow_backend_components.png) #### Data Integration Layer and Zero Copy Tables These data integrations match the external sources mentioned earlier. * **Integration Hub**: will access the REST API data, for the lab it will be a periodic trigger. * **Zero Copy Connector for ERP**: also known as ZCC for ERP will get the cost center master data from SAP. * **Zero Copy Connector for SQL**: also known as ZCC for SQL will connect to the Snowflake data assets for the lab, specifically the Cost Center History, Expenses, and Summary. * **External Content Connector**: will access the indexed documents in SharePoint to enrich decision making and automations for our finance workflow. #### Zero Copy Tables These are pointers to the respective tables coming from either the Cloud Data Warehouse (Snowflake) or ERP (SAP). * **Cost Center Master Data:** taken from the ERP system, known as CSKS from SAP, which can be obtained via BAPI, RFC Table Read, or OData endpoints. Master data does not change frequently in ERP systems so either persistence or zero copy approaches are viable for ServiceNow use cases. * **Cost Center Summary**: aggregated Cost Center History data for the year stored in the Cloud Data Warehouse for reporting purposes. * **Cost Center History**: monthly data taken from the Cloud Data Warehouse with information whether cost centers have historically gone over/under budget or on target. ERP systems normally do not store this type of history or aggregation and is hence stored in a business or data warehouse. **Note** that for the purpose of this exercise, this has been created as a local ServiceNow table but is usually stored in a Cloud Data Warehouse or ERP system in customer environments. * **Expense History**: this can come from an expense management system, ERP system, or Cloud Data Warehouse. For this lab, we are obtaining this from the Cloud Data Warehouse. **Note** that for the purpose of this exercise, this has been created as a local ServiceNow table but is usually stored in a Cloud Data Warehouse or ERP system in customer environments. #### ServiceNow Native Tables These are local tables and are not persisted in any other systems. * **Expense Event Table:** is a scoped table will obtain expense events via Rest API from expense sources such as cloud infrastructure services or even document capture via ServiceNow Lens and Document Intelligence. * **Finance Case Table**: is a standard table in ServiceNow Finance case management (**sn\_spend\_sdc\_service\_request**) where the Flows and Agents will perform updates based on the expense events ingested either in the background or interactively. #### AI & Automation * **Flow and Action:** a scoped Flow and a scoped Action are used to get expense data from a REST API source. * **MCP Server and Client:** MCP or Model Context Protocol is an open standard that lets AI models (LLMs) securely connect to and use external data, tools, and services, acting like a universal adapter to give real-world context beyond their training data for better decision-making and task execution. This will allow Now Assist from ServiceNow act as a data source for clients such as Claude Desktop or act as a client when connecting to MCP-based services (e.g. Snowflake or Notion) feeding data into Flows or Agents. * **RAG - Retrieval Augmented Generation:** Now Assist uses AI Search to retrieve relevant information from your knowledge bases, documentation, and records before generating a response. Rather than relying solely on what the LLM was trained on, it grounds answers in your actual ServiceNow local and integrated data. Think of it as having someone who does the research first, then answers, except it happens in seconds. * **Agent: Proactive Budget Alert Integration Hub Source:** contains RAG and Flows to assess cost center budget history, transactions, and status to create the appropriate Finance Case so budgets can be handled proactively. This is triggered by the Flow and Action that gets Expense Transaction Events from an external REST API source. * **Agent: Over-Budget Case Creator Zero Copy Source**: similar to **Agent: Proactive Budget Alert Integration Hub Source** but does not have a trigger. This agent is created separately to create a lab exercise that focuses on Zero Copy but is similar to **Agent: Proactive Budget Alert Integration Hub Source** with the only difference being the trigger configuration. #### AI Experiences While not part of Workflow Data Fabric, these provide additional data sources for the Flows and AI Agents to use. * **ServiceNow Lens:** an AI-powered feature that uses generative AI to scan, extract, and understand data from on-screen sources like screenshots, emails, or log files, even outside ServiceNow. This can be useful to get Expense Transaction Events coming from receipts which are scanned interactively (i.e. through ServiceNow app or in a browser window). * **Document Intelligence:** an AI-powered solution that uses machine learning, natural language processing (NLP), and computer vision to automatically extract, classify, and process data from various digital and scanned documents, useful for processing Expense Transaction Events coming from multiple documents. ## Overall data flow The diagram below will be further decomposed in the next sections to give you more detail on the inner workings within ServiceNow while, as mentioned earlier, the end user interacts with **Employee Center**, an **MCP Client** (e.g., Claude Code or Desktop), or in slightly more technical scenarios **AI Control Tower**. You can skip the review of the diagram below if you prefer, and head straight into the lab exercises, each of which having its individual (and much more detailed) data flow. {% @mermaid/diagram content="--- title: Workflow Data Fabric Landscape ------------------------------------- graph TB subgraph USER\["User Interaction"] WC\["Workspace"] EC\["Employee Center"] MCP\_C\["MCP Client"] end ``` subgraph SNOW["ServiceNow AI Platform"] direction TB subgraph AI["AI & Automation"] AGENTS["AI Agents"] FLOWS["Flows & Actions"] RAG["RAG via AI Search"] end subgraph INT["Integration Layer"] ZCC["Zero Copy
Connectors"] INTHUB["Integration
Hub"] XCC["External Content
Connector"] MCP_S["MCP"] end subgraph DATA["Data Layer"] ZCT["Zero Copy Tables"] NT["Native Tables"] end subgraph AX["AI Experiences"] LENS["Lens"] DOCINT["Document Intelligence"] end end subgraph EXT["External Systems"] ERP["ERP"] CDW["Cloud Data Warehouse"] DOC["Document Storage"] API["Expense Event API"] EXT_MCP["External MCP"] end USER --> AI MCP_C --> MCP_S AI --> DATA AX --> DATA DATA --> ZCC ZCC --> ERP ZCC --> CDW AI --> INTHUB INTHUB --> API AI --> XCC XCC --> DOC MCP_S --> EXT_MCP %% Styling classDef external fill:#2196F3,stroke:#1565C0,stroke-width:2px,color:#fff classDef user fill:#F5F5F5,stroke:#616161,stroke-width:3px,color:#1a1a1a classDef nowassist fill:#FFB300,stroke:#F57F17,stroke-width:2px,color:#1a1a1a classDef platform fill:#4CAF50,stroke:#2E7D32,stroke-width:2px,color:#fff classDef wdf fill:#9C27B0,stroke:#6A1B9A,stroke-width:2px,color:#fff class ERP,CDW,DOC,API,EXT_MCP external class ZCC,INTHUB,XCC,MCP_S wdf class ZCT wdf class NT platform class AGENTS,RAG nowassist class FLOWS platform class LENS,DOCINT nowassist class WC,EC,MCP_C user" fullWidth="true" %} ``` > **Color Legend:** 🟡 Now Assist | 🟢 Platform | 🟣 Workflow Data Fabric | 🔵 External Systems | ⚪ User Interaction > [📊 View High-Resolution Diagram](https://raw.githubusercontent.com/leojacinto/WorkflowDataFabric-TypeA/main/.gitbook/assets/dataflow_complete_landscape.png) [Take me back to main page](https://servicenow-lf.gitbook.io/the-workflow-data-fabric-loom/readme)