Placeholder image

Agentic AI for SAP

| Chan Jin Park |

AI Copilot

YouTube

Anchor

Spotify

Apple

Episode #229

Introduction

In episode 229 of our SAP on Azure video podcast we talk again about AI and SAP. The last few months have seen an amazing evolution in AI. What started as a simple Bot Framework, evolved to GPT driven Chatbots, to Copilot to users interacting with agents. The latest step – like what we are doing together with SAP on the Joule and Copilot integration – is multi-agent integration. Chan Jin Park, CJ, was on our show twice to talk about his scenarios and demos leveraging Azure OpenAI, Teams and Copilot integrations. I am glad to have him back to show us his latest developments.

Find all the links mentioned here: https://www.saponazurepodcast.de/episode229

Reach out to us for any feedback / questions:

#Microsoft #SAP #Azure #SAPonAzure #AI #AzureAI

Summary created by AI

  • Holger introduced the podcast episode and discussed the evolution of AI, highlighting the transition from simple bot frameworks to multi-agent integrations, particularly in the context of SAP and Microsoft.
    • Podcast Introduction: Holger welcomed the audience to episode 229 of the SAP on Azure Video podcast, mentioning the date and the presence of Robert and Goran to discuss SAP and Microsoft-related topics.
    • AI Evolution: Holger elaborated on the significant advancements in AI over the past few months, starting from simple bot frameworks to GPT-driven chatbots, copilot, and user interactions with agents, culminating in multi-agent integrations.
    • SAP and Microsoft: Holger highlighted the collaboration between SAP and Microsoft, particularly focusing on the integration of drool and copilot, showcasing the latest developments in multi-agent integrations.
  • CJ’s Role and AI Integration:
  • Chan Jin (CJ) introduced themselves and explained their role in integrating SAP with Microsoft AI, focusing on supporting SAP sellers and leveraging AI for various scenarios.
    • CJ Introduction: Chan Jin Park, also known as CJ, introduced themselves as part of Microsoft’s Asian CSAU team, helping SAP sellers and the data and AI team integrate Microsoft AI with SAP data.
    • Role Description: CJ described their role in supporting SAP sellers by assisting with POCs, writing MVPs, and facilitating the adoption of Microsoft AI with SAP data, emphasizing the challenges and opportunities in this integration journey.
    • AI Integration: CJ discussed the evolution of AI integration, mentioning the transition from simple models to more complex scenarios involving copilot and multi-agent systems, and the continuous updates and new services emerging in the AI landscape.
  • Single AI Agent and Multi-Agent Integration:
  • Holger and CJ discussed the transition from single AI agents to multi-agent integrations, emphasizing the benefits and challenges of each approach. CJ highlighted the new Azure AI Agent Services and their capabilities.
    • Single AI Agent: CJ explained the concept of single AI agents, where a single agent handles specific tasks, and discussed the limitations and manual efforts required in this approach.
    • Multi-Agent Integration: CJ highlighted the transition to multi-agent integrations, where multiple agents work together to handle complex tasks, providing more efficient and scalable solutions.
    • Azure AI Agent Services: CJ introduced the new Azure AI Agent Services, which offer managed services for creating and orchestrating multiple AI agents, simplifying the integration process and enhancing capabilities.
    • Benefits and Challenges: Holger and CJ discussed the benefits of multi-agent integrations, such as improved efficiency and scalability, as well as the challenges, including the need for proper orchestration and management of multiple agents.
  • Azure AI Foundry and Agent Playground:
  • CJ demonstrated the Azure AI Foundry and Agent Playground, explaining how these tools simplify the creation and management of AI agents, including knowledge and action components.
    • Azure AI Foundry: CJ introduced the Azure AI Foundry, a platform that provides various AI services, including the new Azure AI Agent Services, to facilitate the development and deployment of AI solutions.
    • Agent Playground: CJ demonstrated the Agent Playground, a tool within the Azure AI Foundry that allows users to create and manage AI agents, specifying knowledge sources and actions to be performed by the agents.
    • Knowledge and Action Components: CJ explained the two main components of AI agents in the Agent Playground: knowledge, which involves data sources and grounding, and action, which involves executing tasks through orchestrated agents.
    • Demo and Features: CJ provided a live demonstration of the Agent Playground, showcasing how to create an agent, specify knowledge sources, and define actions, highlighting the ease of use and powerful capabilities of the tool.
  • Multi-Agent Scenario for Customer Care:
  • CJ presented a multi-agent scenario for customer care, where multiple specialized agents interact to handle customer complaints efficiently. This approach improves response times and accuracy.
    • Scenario Overview: CJ described a multi-agent scenario for customer care, where specialized agents handle different aspects of customer complaints, improving efficiency and accuracy in responses.
    • Specialized Agents: CJ explained the roles of different specialized agents, such as product agents, purchase agents, and customer service agents, and how they interact to resolve customer complaints.
    • Interaction and Orchestration: CJ highlighted the interaction and orchestration between multiple agents, managed by a proxy agent, to ensure seamless handling of customer complaints and efficient resolution.
    • Benefits and Implementation: CJ discussed the benefits of this multi-agent approach, including faster response times and more accurate resolutions, and provided insights into the implementation process using Azure AI services.
  • Future of Multi-Agent Scenarios:
  • Holger and CJ discussed the potential of multi-agent scenarios in various contexts, including supply chain management and employee onboarding. They emphasized the ongoing development and future possibilities of these integrations.
    • Potential Applications: Holger and CJ discussed potential applications of multi-agent scenarios in various contexts, such as supply chain management and employee onboarding, highlighting the versatility and scalability of this approach.
    • Ongoing Development: They emphasized the ongoing development of multi-agent integrations, with continuous improvements and new features being added to enhance capabilities and address emerging challenges.
    • Future Possibilities: Holger and CJ explored future possibilities of multi-agent scenarios, envisioning more sophisticated and powerful integrations that can handle complex tasks and provide significant value to businesses.
    • Customer Adoption: They discussed the current state of customer adoption, noting that while some customers are already leveraging these integrations, many are still in the early stages of exploring and experimenting with multi-agent scenarios.

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