Karl-Johan Spiik
Microsoft MVP Power Automate, AI Ambassador, Power Platform Architect, Author, Leadership hacker | Solution Architect @ Digia
Kaarina, Finland
Actions
Karl-Johan has done 18 years projects with Microsoft technologies. Past six years he has been working with the Power Platform, particularly with Power Automate, Power Apps and Copilot Studio, and have extended its capabilities with Microsoft 365 and Dynamics 365 products as well as Azure resources. Karl-Johan has completed about 100 projects with various industries and organization sizes as role of Architect, Developer and Scrum Master. Karl-Johan has written four books about Community-Led Orientation and masters shared leadership theory and practice as well.
Badges
Area of Expertise
Topics
From Flows to Autonomy: Building Intelligent Agents with Copilot Studio
As developers, we’re constantly navigating the evolving landscape of automation and AI. In this session, I’ll share my journey of building autonomic agents using Copilot Studio, exploring the real-world challenges and decisions that shaped the process. You’ll learn how to distinguish when to use AI orchestration for AI-driven decision-making versus Agent Flows for traditional cloud-based workflows.
We’ll dive into:
• The architecture and design of an autonomic agent
• Key lessons from integrating AI capabilities into business processes
• Practical criteria for choosing between Agent Flows and AI prompts
• Performance, efficiency, and maintainability trade-offs
Whether you're just starting with Copilot Studio or looking to scale your AI automation, this session will equip you with the insights and patterns to build smarter, more autonomous solutions.
Beyond Resumes: Developing Autonomous Agent for Efficient Job Application Automation
Job application processes often fall short of meeting the needs of both applicants and employers. This session will explore the development and implementation of autonomous agents designed to automate and enhance job application workflows, making the process more efficient and less stressful for everyone involved.
Imagine an agent that preps the applicant beforehand and provides feedback on their application. The autonomous agent could read the application description and instructions from the employer, offering feedback to the applicant before the application is even submitted. It would then rate the applications based on how well they meet the criteria. Additionally, the agent saves applications, CVs, and other materials into a repository, organizing them for the employer.
For job-related questions, a chatbot could read from a job position database. If there is no answer or the applicant is not satisfied with the response, the agent would ask the person responsible, add the answer to the database, and provide it to the applicant. This way, the person responsible only needs to answer certain questions once.
The solution leverages Copilot Studio, Autonomous Agents, Dataverse databases, Teams actions, email reading and replies, AI orchestration, and more. This solution can be adapted to any situation where a responsible person needs to manage many written applications meeting specific criteria and answer questions. Once built, it can be generalized to handle multiple topics and applications across various channels.
Key Takeaways:
- Building Autonomous Agents: Learn the steps to develop and implement autonomous agents.
- Database Integration: Understand how to read database data to identify application specifications that need attention.
- Effective Communication: Discover how to ask questions from all participants in the process and record their replies in the correct context.
- Multichannel Service: Explore how to offer multichannel services to trigger autonomous agents.
- Applicant Feedback: Learn how to provide valuable feedback to applicants, helping them fine-tune their applications to meet requirements.
Beyond Linear Workflows: Embracing Community-Led Autonomy with AI Agents
What happens when we stop thinking of AI agents as tools—and start thinking of them as teammates? In this session, we’ll explore the shift from traditional, linear workflows to community-led, autonomous systems where agents and humans collaborate in shared leadership models.
You’ll learn:
• The philosophical and practical differences between delegating to humans vs. agents
• How autonomous agents can take initiative, adapt, and contribute to team goals
• The role of community-led orientation in designing agents that align with collective values and goals
• Real-world patterns for task delegation, feedback loops, and trust-building with AI agents
We’ll also discuss how to design agents that are not just reactive, but proactive—capable of participating in a team’s mission, not just executing instructions. This session is ideal for developers, designers, and leaders interested in the future of collaborative AI.
When Azure DevOps Pipelines Gives You Lemons: Use Power Platform Pipelines
Usual scenario is that there is hurry and creating pipelines are dropped away from project early in sales phase or at least in rushed UAT testing and production release. Finally, the DevOps pipelines are done, but consultants are too afraid to use them or cannot find the error messages. It would be better to have “some” pipelines, but IT and technical persons belittles Power Platform pipelines. This session will provide exploration of both Azure DevOps pipelines and Power Platform pipelines.
Main thing is that if you have only one development environment, DevOps pipelines features will never be used since there is always one package going from one environment to another. This session demonstrates scenarios where there could be multiple development environments and DevOps pipelines compared to more streamlined scenario, where there is only one dev, test and prod environment.
Session outcome will present the corner stones when to select Power Platform or DevOps pipelines. There is no point of shoot a fly with a cannon just because preparing to possible integration and automated testing scenarios on some other phase which is not actually decided yet. Power Platform pipelines have it’s advantages, Copilot notes from the solutions and user friendly user interface. Coming Dataverse and Git integration will also remove doubts that source codes won’t we available with Power Point pipelines.
Key Takeaways:
• Understand the core functionalities and features of Azure DevOps pipelines and Power Platform pipelines.
• Learn the advantages and disadvantages with different amount of development environments.
• Gain practical experience through hands-on demonstrations and setup guidance.
• Discover best practices and lessons learned from real-world case studies.
Join us for this insightful session to deepen your understanding of Azure DevOps and Power Platform pipelines and learn how to choose and implement the best solutions for your organization's needs.
Mastering Power Automate Flows: Advanced Techniques and Best Practices
Developing efficient and reliable Power Automate flows is crucial for business automation and productivity. This session addresses the challenges and solutions for creating advanced Power Automate flows.
Key Takeaways:
• Master iterative development for easier debugging
• Implement child flows and 'run only' user settings
• Configure dynamic responses in child flows
• Implement robust logging and error handling strategies
• Handle errors with pre-termination responses
• Use trigger actions and Terminate functions effectively
• Manage "Save and close" in model-driven apps
• Bypass offline profiles for uninterrupted automation
• Set up custom spam prevention settings
• Leverage AI models like GPT for automation
• Manage approval flows in Teams or out-of-the-box
This advanced session on Power Automate flows is aimed at experienced users. The workshop will cover:
• Iterative development and debugging techniques
• Child flows and 'run only' user settings
• Logging and error handling strategies
• AI integration for prompt data operations
• Approval flows in Teams channels
Participants will learn methods to build and iterate Power Automate flows for simplified debugging, break down complex workflows into manageable parts, and understand the setup of child flows and the concept of 'run only' users for secure execution of tasks. The session will also cover configuring child flows to respond before the parent flow concludes, enabling dynamic automation solutions.
The session will include interactive demonstrations, real-world examples, and opportunities for audience participation to ensure a comprehensive understanding of the topics covered.
The session will include interactive demonstrations, real-world examples, and opportunities for audience participation to ensure a comprehensive understanding of the topics covered.
Please note that Sessionize is not responsible for the accuracy or validity of the data provided by speakers. If you suspect this profile to be fake or spam, please let us know.
Jump to top