Posts

Copilot vs Agent: The Simple Difference Everyone Gets Wrong

If you use AI at work, you have heard two words a hundred times: Copilot and Agent. Most people think they are the same. They are not. Here is the simplest explanation you will ever read. The One Line Difference Copilot helps YOU do the work. Agent DOES the work FOR you. That's it. Let's break it down. What is a Copilot? Think of a car. You are the driver. Copilot is the person sitting next to you, helping with directions.         You are in control        You ask, it answers in 5 seconds          It works inside one app at a time          It always waits for you Real Example: You open Outlook and say: "Summarize this long email from my client." Copilot summarizes it. Done. It stops. You are still the one doing the work. Copilot just makes you 3x faster. What is an Agent? Now think of an employee you hire. You give them a goal. They make a plan, use 10 differ...
Post a Comment
Read more

Building Stateful Agentic AI Workflows with Azure Standard Logic Apps

Introduction: What are Agentic AI Workflows in Azure Logic Apps? Azure Standard Logic Apps are no longer just about "connecting APIs". With AI, they can now act like agents - making decisions, learning, and running workflows with minimal human input. In this post, we’ll break down how to build Stateful Agentic AI Workflows using Azure Standard Logic Apps, when to use them, and the pros/cons you should know before going live.   1. Two Types of Agentic AI in Logic Apps When building an AI agent in Logic Apps, you have 2 main patterns: A. Autonomous Agent 🤖 Think " set it and forget it ". What it does: Operate independently with minimal human intervention. Make decisions based on goals + data.  Best for: Repetitive, well-defined tasks like data processing, scheduled jobs, background automation. Key traits: Self-directed, no constant prompts needed, clear objective. B. Conversational Agent 💬 Think " ChatGPT inside your workflow ". What it does: Talks with...
Post a Comment
Read more

Build a GraphQL API with Synthetic GraphQL in Azure APIM!

Image
We've already nailed PassThrough GraphQL in APIM 🚀( https://tech-sakshimittal.blogspot.com/ ). Now, let's dive into Synthetic GraphQL💪 Why Synthetic GraphQL? 🤔 It's perfect when you want to: - Aggregate data from multiple sources (REST APIs, Azure SQL, Azure CosmosDB) - Simplify complex backend interactions - Give clients flexible query power 😎 - Transform or shape existing REST APIs without changing them - Expose a more query-friendly interface to clients 👉  Data Sources In this demo, we've set up 3 data sources whose responses will be mapped to GraphQL schema fields. 1. Azure SQL 2. REST API 3. Azure CosmosDB  👉 DataSource1: Azure SQL [dbo].[bookInfo] In Azure SQL database, a table is created ' [dbo].[bookInfo]'. Here is the data: 👉  DataSource 2: BookStore REST API        These are the REST API endpoints we'll connect to GraphQL via APIM . GET /books → Fetch all books 📚(list of all books)        This endpoint returns...
Post a Comment
Read more

Exposing GraphQL APIs Securely with Azure APIM: A Step-by-Step Guide🚀

Image
  Are you looking to expose GraphQL APIs securely? Azure API Management (APIM) provides robust support for GraphQL, making it easier to manage and secure your APIs.   In APIM, there are 2 options available to expose GraphQL api.         Pass-Through GraphQL: Proxy an existing GraphQL service (hides original service details)          Synthetic GraphQL: Build a GraphQL API on top of REST APIs/other services This article covers Pass-Through GraphQL. Step 1: Create a GraphQL API in APIM - Azure portal > APIM > APIs > + Add API > GraphQL Step 2: Import GraphQL Schema              Provide GraphQL endpoint, API URL suffix, display name             APIM auto-fetches schema (if endpoint is accessible). If it does not do auto fetches, choose to import a schema file(*.graphql). Click on Create button.  Please Note: In case of endpoint. If a...
Post a Comment
Read more