Mastering AI in Power BI

A full end-to-end masterclass on AI in Power BI. During this training day, we’ll cover every corner of AI in Power BI, from preparing data for AI to bringing your Power BI AI assets to life with MCP.

In these times, AI literacy is both an essential career skill and a powerful value-add for organizations. The ability to quickly and easily draw insights and get answers from our data is as important as ever. Furthermore, being able to interact with our data through conversational communication is increasingly in demand.

The training day will be a mix of presentation-style teaching, interactive exercises, and workshop labs. After attending this training day, you’ll be able to set up everything around AI in Power BI end-to-end. You’ll be up to date on the newest tools for AI-powered BI—ready to develop solutions or advise clients and colleagues on how to use AI with Power BI.

Module 1: Data Prep for AI – with AI

In this module, we’ll explore how to get your data ready for AI usage. We’ll cover semantic modeling best practices, context- and prompt engineering, and how to leverage AI itself to fine-tune semantic models for AI usage.

Module 2: AI-assisted Power BI development

In this module, we'll explore the various options for improving your development productivity with the available AI toolbox. We'll cover how the built-in Copilot may help improve low-code workflows and move on to full-fledged AI coding agents. At this stage, participants will learn how to make AI agents develop whole user stories with no human involvement.

Module 3: Data Agents

The third module covers everything about using semantic models as assets for AI agents. In this module, you'll learn how to set up, configure, and refine data agents so that your semantic model data is readily available for "chat with your data" scenarios, as well as for use outside of Power BI.

Module 4: External solutions

In the final module, we explore how Power BI data can be used with AI tools outside of Power BI itself. We'll cover how to use MCP to create custom AI agents and how to use the native MCPs for quick results. Furthermore, we'll explore how to integrate your data agents and semantic models with the rest of the Microsoft AI ecosystem, using AI Foundry and Copilot Studio. Finally, we'll cover how to move beyond the boundaries of Microsoft and use the data in cloud-agnostic use cases, using today's most talked-about AI tools.

Advanced Agent Skills for Power BI

From the moment you opened VS Code, Claude Code, or a similar tool, and perhaps connected to a Power BI MCP, you had already leaped miles ahead of the curve. If you also expanded the capabilities of those agents with Agent Skills, I'm sure you haven't looked back since.

Agent Skills, at the surface, are extremely impactful but also very simple technically, being mostly just text files and folders. But did you ever get a chance to look beneath that surface?

In this session, we'll cover the hidden depths of what you can achieve with Agent Skills, the open standard for extending agent capabilities. First, we look at how skills are discovered, loaded, and interpreted by agents. Second, we explore how skills can be much more than just instructions with scripts, hooks, and more. Third, how skills can be bundled with sub-agents to run everything more parallelized, or how to make the skill itself run as a sub-agent. Finally, we discuss how to add self-learning loops into skills so that they evolve with your project and document your work as you go.

Topics:
- Power BI modeling with AI and MCPs
- Agent Skills, hooks, and scripts
- Sub-agents and self-learning agents

After attending this session, you will know everything there is to know about Agent Skills for Power BI, and leave with a refined AI toolbox, able to build even more ambitious agentic workflows for Power BI.

Organizing Your Data Lake

In today's data-driven landscape, creating an efficient data structure for your data lake has become an indispensable part of any successful data management strategy. It's no longer sufficient to merely land your data in a data lake - it's critical to plan your data structure ahead of time to maximize utility and performance.

In this enlightening session, we will delve into the importance of pre-emptive planning for your data structure. We will explore how a well-thought-out data structure can dramatically enhance the efficiency of security, partitioning, and processing strategies.

Attendees will gain insights into best practices for structuring data and will learn techniques to leverage these structures for optimal performance. This session aims to equip you with the knowledge to harness the power of your data lake effectively

Chat with your data

In the era of AI, the ability to quickly and easily draw insights and answers from our data, is as important than ever, and being able to ask those questions in plain human language.

In this session we'll uncover the possibilities of combining Microsoft Fabric with OpenAI, to enable scenarios like chatting with our data. We'll touch upon topics like the OpenAI API, prompt engineering, prompt flow and how to data model our data for easy consumption by the AI.

Topics:
- Open AI API
- Prompt engineering in prompt flow
- Data modelling for GPT

After attending this session, you'll leave with an understanding of how to harness the power of Microsoft Fabric and OpenAI to revolutionize the way you interact with data.

Tuning Agentic Coding for Fabric

Coding agents are getting better by the week. But out of the box, they don’t understand your data platform, your standards, or your architecture.

In this session, we’ll explore how to fine-tune coding agents to work effectively with Microsoft Fabric. We’ll look at how to guide agents with context from your Lakehouse, medallion architecture, naming conventions, and governance principles, so they generate code that actually fits your platform.

In this session, we focus on how to shape and control coding agents when working with Microsoft Fabric. We’ll cover tools like GitHub Copilot, Claude Code, VS Code extensions, and compare different coding agents and workflows.

You’ll learn how to define agent skills, agent instructions, and build custom agents tailored to your data platform, and how to distribute it and share it seamlessly with anyone using agents in your Fabric tenant. We’ll also explore MCPs and how to connect agents safely and effectively to external context and services.

Vibe-coding for the data professional

Agentic coding has changed how we write code, but it has also changed who gets to write it and how quickly ideas turn into working solutions.

In this session, we’ll explore agentic engineering, popularly called vibe-coding, a flexible, extremely fast approach to building prototypes, proofs-of-concept, or even mature solutions.

We go from easy practical examples related to Microsoft Fabric that anyone could vibe-code in an afternoon, and moving on to advanced features, like Microsoft auth with MSAL, Utilizing Fabric API’s in application back-ends, creating browser extensions and even vibe-coding locally installed desktop software.

Obviously, we’ll also discuss the limitations and risks of vibe-coding, and its ideal use-cases.

After this session, I promise you that you won’t frown when you hear the term vibe-coding anymore, but I can’t promise you that you won’t spend the next few nights restlessly vibe-coding yourself.

Agentic AI in Fabric

In the era of AI, the ability to quickly and easily draw insights and answers from our data, is as important than ever, and being able to ask those questions in plain human language.

In this session we'll uncover the possibilities of leveraging AI in Microsoft Fabric to achieve even more. We'll touch upon topics like Fabric data agents, data agent SDK, Microsoft Foundry, Agentic coding with Github Copilot, and vibe-coding extensions to Fabric Data Agents and Fabric in general.

Topics:
- Using AI to prep data for AI
- Fabric Data Agents
- Advanced Data Agent code-only features
- Microsoft Foundry
- Leveraging data agents outside of Fabric
- Vibe-coding data apps

After attending this session, you’ll be able to leverage your data with AI agents inside and outside of Fabric, and you be up to date on the newest tool for AI-powered BI, ready to develop solutions or advice clients and colleagues on how to use AI with Fabric.

Navigating Data Modeling in Direct Lake

In this session, we'll dive into the Direct Lake, the new Power BI storage mode that seamlessly blends DirectQuery's real-time capabilities and Import mode's performance.

We unravel how this promising technology impacts traditional data modeling approaches, posing essential questions on star schemas, surrogate keys, SCDs, and other classical dimensional modeling techniques.

This session equips you with a deeper understanding of the Direct Lake landscape and its implications for data modeling, empowering you to craft performant data models.

The key areas we'll explore are:

- Overview of Direct Lake
- Impact of Direct Lake on data modeling techniques
- Redefining data transformation strategies

We unravel how this promising technology impacts traditional data modeling approaches, posing important questions on star schemas, surrogate keys, SCDs and other classical dimensional modelling techniques.

Next-gen dataflows in Microsoft Fabric

Have you also dreamt about using Power BI dataflows as a data integration tool beyond the Power BI service? Perhaps instead, you've experienced the dread of working in a dataflow with too much data or too many objects, and experienced the sluggishness that can come with it?

Don't wait any longer - Dataflow Generation 2 is here!

In this session, we assess if this next generation of dataflow will end all our problems and be everything we dreamed of. We'll cover the differences between the new generation and its predecessor and uncover new features and UI changes. We also cover any limitations and assess if Generation 2 truly lives up to its intended purpose.

After attending this session, you will know your way around Dataflows Generation 2 and its underlying spark engine. You'll understand their purpose, use cases, and implications for future data platform architectures.

The key areas we'll explore are:

- The underlying Spark engine
- Data destinations & Pipeline integration
- New no-code UI
- The role of dataflows gen 2 in a modern data platform
- Capacity and cost implications

Data Contracts: The Fabric of Data Quality

Data quality is the heart of any successful data-driven project. Imagine defining a set of standards data sources owners need to meet, and then track their adherence to these standards. Inspired by API best practices, Data Contracts offer a sturdy system to establish and handle data agreements.

In this session, we'll explore a pragmatic approach to manage, monitor, and react to data quality deviations using a blend of Microsoft Fabric tools, Power BI and GitHub.

Whether you're new to Data Contracts or seeking to embed them within Microsoft Fabric, this session offers a hands-on guide to using this powerful data management tool in your Power BI and Microsoft Fabric environment. Join us, and learn to keep your data clean, consistent, and reliable.

The key areas we'll explore are:

- Data contracts in Microsoft Fabric
- Data contract repository in GitHub
- Developing a data validation engine using Notebooks
- Visualizing data contracts and quality checks in Power BI

Object-Level Security: Control access to entire columns and tables

Is row-level security just not enough for you? Maybe you wish that you could limit user access to entire columns or even tables?

You can now! Microsoft recently introduced Object Level Security (OLS) to Power BI.

During this session you will learn about the proper use-cases for using Object-Level Security as well as what pitfalls and problems to avoid while we go through how to set up Object-Level Security in an actual datamodel in Power BI.

After this session you will be able to:
- Set up Object-Level Security for your own data models, to restrict access to columns or tables
- Avoid common pitfalls and problems caused by Object-Level-Security
- Assess if Object-Level Security is the right solution for you user need

Designing and managing a cost-effective data platform in Azure Synapse Analytics

Azure Synapse Analytics is a powerful data platform, but it can also be expensive if you don't know what you're doing. In this session, we will go through the different components of Azure Synapse Analytics and discuss how to design a cost-effective data platform.

We will cover topics such as choosing the right pricing tier, optimizing data storage and processing, and leveraging built-in cost management features. We will also discuss how to optimize your data platform for cost efficiency by using features such as serverless compute, pre purchase compute and reserved capacity.

Attendees will leave with a better understanding of how to design and manage their Azure Synapse Analytics platform for cost efficiency and how to design a cost-effective data platform that meets your organization's needs.
Topics we will cover:

- Pricing Models for compute Resources in Azure Synapse Analytics
- Storage Types and Tiers
- Pricing Models
- Optimizing Cost with Resource-Scaling
- Using serverless compute and reserved capacity options for cost savings

Supercharge Power BI with Azure Synapse Analytics

Power BI is incredibly powerful at distributing and sharing data insights, through visualizations, across an entire organization. The democratization of Business Intelligence that Power BI brought with it is unparalleled. But Power BI on its own has its limitations and will often fall short, on its own, when designing end-to-end analytic platforms for enterprises. That’s where Synapse comes into the picture-for organizations, small and large, their current Power BI deployment can be supercharged with the capabilities of Azure Synapse.

In this session we will cover how to architect an end-to-end enterpise analytics platform so that we utilize the best of both worlds. The extremely versatile ETL capabilities of Synapse and the powerful data modelling in Power BI.

After this session you will leave with a much better understanding of how Azure Synapse Analytics can supercharge your current Power BI.

Datamarts – The good, the bad, and the ugly

You may think of Datamarts as this brand new thing in Power BI – and it is. But it is also a concept coined already in the 70s. But enough history, you are probably here to learn how you can utilize the new Power BI Datamarts in your organization.

In this session, we’ll cover everything A-Z within Power BI Datamarts. What you can do and how you can do it. We also take it a step further and discuss the architectural pros and cons of using Power BI Datamarts over alternative tools such as SQL Servers, Data Lakes and Azure Synapse Analytics.

After attending this session you’ll be an expert on Power BI Datamarts, knowing every ins and outs of the feature, sprinkled with knowledge about overall Business Intelligence architecture principles. All in all, you should be capable of evaluating if Power BI Datamarts deserves to play a central role in your BI data architecture, now or in the future.

Topics:
- What is a Datamart
- A-Z walkthrough of Power BI Data Marts
- Strength and pitfalls of Power BI Datamarts
- Business Intelligence Architecture
- Using Power BI Datamarts with Azure Synapse Analytics

Advanced Power BI refresh with ADF/Synapse

Want to refresh your Power BI Datasets in ways the service doesn't allow you to, out of the box? There are a few reasons why the UI in Power BI isn't sufficient for optimal dataset refreshing in Power BI.

In this session, you will tag along the journey of working towards a solution to this problem. We go through the journey of a real-life case where a Power BI set-up ended up needing advanced refresh handling.

After this session, you will be able to conceptualize and implement different Power BI refresh scheduling patterns in ADF/Synapse. You’ll learn what limitations in Power BI you can overcome with an “external” orchestration tool as well as the different building blocks available. Finally, we will go through an example implementation as well as a step-by-step demo of how you can set this up yourself.

Topics:
- Why use a different orchestration tool than Power BI itself?
- Exploring the toolbox – what options do we have?
- Example implementation: Asynchronous Power BI refresh schedules orchestrated by ADF
- Demo: How to set it up in practice

Loading data with Rest API in Power Query

Do you want to load data into Power BI with REST APIs? REST APIs are everywhere. They are a great way to get data from one system to another, but they can be difficult to work with if you don't know where to start.

In this session, you will learn how to use Postman to explore a REST API. You will learn how to find the information you need in the API documentation, and how to use that information to create a Power Query function that loads the data into Power BI using the built-in web connector.

After attending this session you will be able to find and use API documentation, authenticate with an API, and prototype API calls in Postman. You will also have knowledge about the use of the Power Query REST API connector to load data from REST APIs into Power BI.

Topics:
- REST API
- API Documentation
- API Authentication
- Prototyping in Postman
- Power Query

When data lies - typical patterns for manipulating data

Data should be impactful, and to be that, it needs to influence decisions or result in changes. But how much harm could such impact cause if said data is erroneous or straight-up manipulated?

How can data be manipulated? That is precisely what we will explore in this white-hat-inspired session. We will put ourselves in the shoes of a data villain and explore ways to misrepresent reality and make our data lie to the best of our ability.

In this, part humorous, part delicate session, you will learn to detect malicious patterns in data that could suggest foul play.

Topics:
- Data Stories
- Data Manipulation
- Data & Analysis integrity

Complex access control with Row-Level Security

How do you manage data access with hundreds of users spread across departments, countries, or even companies while using the same Power BI dataset across those entities?

In this session, you will tag along the journey of working towards a solution to this problem. We will touch upon topics like automated workflows with Power Automate, Active Directory Security Groups, Decoupled Datasets and Security Tables, all combined with Row-Level Security to bring them all together.

You will be introduced to various advanced methods for handling data access in Power BI so that you will also be able to design highly flexible and dynamic security setups. The whole ordeal will be sprinkled with my experiences, failures, and successes in implementing such a setup in a multinational company.

Topics:
- Row-Level Security
- Workflow Automation (Power Automate)
- Decoupled Datasets
- Active Directory Security Groups
- Security Tables

A deep dive into Direct Lake

The new Direct Lake storage mode in Power BI promises to revolutionize data handling by blending the real-time nature of DirectQuery and the high performance of Import mode. Is this all just hype, or does Direct Lake truly deliver on these claims?

In this session, we will put Direct Lake under the microscope, conducting comprehensive performance testing across varied data sets and query types. We aim to stress test Direct Lake, uncovering where it meets expectations, where it excels, and where it has room for improvement.

The key areas we'll explore are:

• Direct Lake: Understanding what we know about the engine
• Analyzing performance across different scenarios
• Strengths and weaknesses
• Advice for data modelling for Direct Lake

By the end of this session, you'll have a well-rounded understanding of Direct Lake's capabilities, potential for your Power BI data modelling projects.

the AI workforce revolution

Your biggest competitor isn’t who you think. It’s a 3-person startup with AI agents doing the work of 50. While you debate AI governance, young founders are automating entire departments with tools that cost less than your car's lease.

The adoption gap is real: startups ship in weeks what takes enterprises 18 months. Drawing on insights from real-life cases and technology adoption research, this session explains why most companies are trapped in AI’s “disappointment phase” while missing the revolution happening in production.

You’ll leave with a survival blueprint—because in 24 months, every industry will be divided into AI-native winners and everyone else.