Placeholder image

Copilot & AI Agents for SAP

| Amit Lal |

AI Agentic

YouTube

Anchor

Spotify

Apple

Episode #238

Introduction

In episode 238 of our SAP on Azure video podcast we talk about next wave of AI. Last week we heard from CJ about a cool usage for the office of the CFO, combining Semantic Kernel, Azure AI Agents and Code Interpreter. This time we have also a known guest with us: Amit Lal. Amit was one of the first to publish SAP related apps to ChatGPT, he released several Agents for Microsoft Copilot and he always has amazing demos to show for our customers. There is now even a nice action figure of Amit available – well, at least virtual.

Find all the links mentioned here: https://www.saponazurepodcast.de/episode238

Reach out to us for any feedback / questions:

#Microsoft #SAP #Azure #SAPonAzure #AI #Agentic

Summary created by AI

  • Holger introduced Amit Lal, a Principal Technical Specialist with Microsoft Strategic Partnership Team, who has extensive experience in SAP analytics, AI, and cloud migration. Amit runs an AI lab focused on practical solutions for SAP technical business pain areas.
    • Introduction: Amit Lal introduced himself as a Principal Technical Specialist with Microsoft Strategic Partnership Team, emphasizing his five years of experience in SAP analytics, AI, and massive cloud migration. He highlighted his work with SAP and major consulting firms, tackling complex technology challenges and large-scale migrations.
    • AI Lab: Amit mentioned running an AI lab that focuses on creating practical solutions for SAP technical business pain areas. He shared that he has 13,000 followers who support his work in AI innovations.
  • AI Agent Demo:
  • Amit demonstrated his AI agent, which collects and sends information from various sources, including news channels, to keep users updated. He offered to share the open-source code for building similar agents.
    • AI Agent: Amit demonstrated his AI agent, which collects information from various sources, including news channels, and sends updates to users. He highlighted the agent’s ability to provide timely and relevant information.
    • Open Source: Amit offered to share the open-source code for building similar AI agents, making it accessible for others to replicate and use in their own environments.
  • Copilot and AI Agents:
  • Amit discussed the use of Copilot and AI agents in SAP and non-SAP interfaces, highlighting their capabilities in handling business processes and providing deep analysis and research.
    • Copilot Usage: Amit explained how Copilot and AI agents are used in both SAP and non-SAP interfaces to handle various business processes. He emphasized their role in providing deep analysis and research capabilities.
    • Daily Use: Amit shared that he uses Copilot and AI agents daily, particularly for tasks such as deep analysis and research, showcasing their practical applications in real-world scenarios.
  • Agentic Framework:
  • Amit explained the four pillars of the agentic framework: memory, planning, action, and tools. He emphasized the importance of these components in building effective AI agents.
    • Four Pillars: Amit detailed the four pillars of the agentic framework: memory, planning, action, and tools. He explained that these components are essential for building effective AI agents.
    • Memory: Amit highlighted the addition of memory to the agentic framework, which allows agents to retain information and maintain a chain of thought, enhancing their functionality.
  • Multi-Agent Scenarios:
  • Amit showcased multi-agent scenarios where different agents handle specific tasks, such as finance, Salesforce, and SAP analytics, with a coordinator agent collecting and passing information to humans for decision-making.
    • Specific Tasks: Amit described how different agents handle specific tasks, such as finance, Salesforce, and SAP analytics. Each agent is specialized in its area, ensuring efficient task management.
    • Coordinator Agent: Amit explained the role of the coordinator agent, which collects information from various specialized agents and passes it to humans for decision-making, acting as a digital employee.
  • AI Use Cases:
  • Amit presented various AI use cases, differentiating between low code and pro code solutions, and highlighted the capabilities of Microsoft Copilot Studio and Azure AI Foundry.
    • Low Code Solutions: Amit discussed low code solutions, such as Microsoft Copilot Studio, which allow business users and citizen developers to quickly create applications and solutions without extensive coding knowledge.
    • Pro Code Solutions: Amit highlighted pro code solutions, such as Azure AI Foundry, which are designed for more complex and robust applications requiring advanced technical skills and machine learning capabilities.
  • API Monitoring Agent:
  • Amit demonstrated an API monitoring agent that triggers alerts and assembles team members in a Teams channel when an API failure occurs, saving time and effort in addressing the issue.
    • API Monitoring: Amit demonstrated an API monitoring agent that continuously monitors APIs and triggers alerts when a failure occurs. This agent sends notifications and assembles the relevant team members in a Teams channel to address the issue promptly.
    • Time-Saving: Amit emphasized that the API monitoring agent saves significant time and effort by automatically gathering the necessary team members and initiating the resolution process without manual intervention.
  • Procurement and Supply Chain Agents:
  • Amit showcased agents for procurement and supply chain management, which help determine the best sourcing options by analyzing tariff information, supply chain disruptions, and real-time data from SAP systems.
    • Tariff Analysis: Amit demonstrated agents that analyze tariff information and supply chain disruptions to determine the best sourcing options. These agents use real-time data from SAP systems to provide accurate and timely recommendations.
    • Decision-Making: Amit explained how these agents assist in decision-making by evaluating various factors such as lead time, risk, and cost, helping companies make informed choices about their procurement and supply chain strategies.
  • Human in the Loop:
  • Amit highlighted the importance of human intervention in AI-driven decision-making processes, showcasing a voting system where employees and the CFO office can approve or reject AI-generated options.
    • Voting System: Amit showcased a voting system where employees and the CFO office can approve or reject AI-generated options. This system ensures that human judgment is incorporated into the decision-making process.
    • Human Intervention: Amit emphasized the necessity of human intervention in AI-driven processes, stating that while AI can provide recommendations, final decisions should involve human oversight to ensure accuracy and accountability.
  • Financial Strategy Agents:
  • Amit discussed financial strategy agents that analyze demand and supply situations, helping determine the best strategies for product launches and market expansion.
  • GitHub Copilot for Fiori Apps:
  • Amit demonstrated the use of GitHub Copilot to create Fiori apps based on whiteboard images, showcasing the ease and efficiency of building applications with AI assistance.
  • Learning Resources:
  • Amit provided information on self-learning labs and interactive workshops available on his GitHub repo, encouraging partners and enterprise customers to start learning and leveraging AI capabilities.

Links

Video index

Based on and using Hugo Cookbook Theme, Font Awesome Icons and Bulma CSS. | Thanks to Recipes4AJ to get me started. | Contact me via