Kurt Buhler

It's now possible to use AI and agents to not just chat with your data, but build semantic models, dashboards, and end-to-end data solutions. However, this is an area that's advancing rapidly; it's overwhelming for everyone and difficult to know where to get started. In this session, we'll give you a beginners introduction to using agents for creation experiences so that they can save you time and enable you to do more. You can expect: - The difference between conversational BI with data agents and Copilot versus agentic development. - An introduction to coding agents and the foundational basics of how they work and why they can be valuable (and dangerous if used irresponsibly). - Demonstrations of how to use coding agents to work with Power BI items like semantic models and reports, as well as to interface with Fabric. - An explanation of the core "building blocks" of effectively using agents and AI, with focus on how to choose the right tools, and curate the right context. - A step-by-step walkthrough of how to get started using agents to help you build in Power BI and Fabric. - Discussions on the future of business intelligence, knowledge work, and more given the advancing capabilities of agents For this session: You do not need any experience in using AI or agents, but we recommend a basic familiarity with Claude Code, the PBIP format, Git/GitHub and source control (as well as of course Power BI and the Fabric platform) By the end of this session, you will have a solid foundational understanding of agentic development and its value, how you can continue learning and experimenting yourself, and where the future of BI and data analytics might be headed.

Power BI reports are well known and widely used, but sometimes you need more control over layout and interactivity than a standard report page allows. Web based reporting using D3.js or Vega is a mature solution in this area that even predates Power BI, but has been too complex for most teas, requiring specialized skills and resources. However, agents make this faster, easier, and more cost effective. This means that you can build custom, bespoke dashboards with powerful visualizations catered to your specific business needs. This session walks through how to build custom web-based dashboards in Microsoft Fabric using coding agents and established libraries. We'll connect to semantic models, query the data, and render it in fully custom HTML interfaces that sit on top of the same trusted models your organization already maintains. You'll see what becomes possible when you're not boxed in by standard visual constraints, formatting limits, or canvas dimensions. In this session, we'll - Explain what a web-based dashboard is and how it compares to a report from an SSBI tool. - Explain how to create one with a coding agent, and how it works. - Discuss the benefits and advantages as well as any drawbacks and risks for those thinking to use this approach. - Provide you general tips for using AI and agents to build visualizations and dashboards. By the end you'll have a clear picture of when reaching for code and agents makes sense, and what the actual implementation path looks like if you decide to try it. Familiarity with frontend concepts, coding agents like Claude Code, and visualization in Fabric will help you get the most out of this session.

Microsoft Fabric brings many new features that change the way that you build and use your Power BI semantic models (formerly datasets). To get the most from Power BI, you should know what these features are, as well as how and when to use them.

Microsoft Fabric brings many new features that change the way that you build and use your Power BI semantic models (formerly datasets). To get the most from Power BI, you should know what these features are, as well as how and when to use them.

Microsoft Fabric brings many new features that will change the way that you build and use your Power BI semantic models (formerly datasets). To get the most from Power BI, you should know what these features are, as well as how and when to use them. In this talk, I’ll discuss: - What a Power BI semantic model is - Why it's so important in Fabric - Three example scenarios to illustrate how different teams choose and use the new Fabric features to build and use their semantic models to address pains and problems that they have in Power BI today. I'll also highlight the considerations these new features bring. In this talk, it's assumed that you've already heard the basics of Fabric features like Direct Lake storage mode, semantic link in notebooks, or Git integration. I'll focus on the "how and why" of these features, over the "what". By the end of this talk, you'll have a better understanding of how you might approach these new features for the semantic models you manage today, or will make in the future.

Many reporting requests have a common starting point - existing reports. These might be Excel workbooks, large PowerPoint slide decks, or bundles of reports scattered across other technologies. For one reason or another, these reports need to be re-made, re-designed or migrated over to Power BI. “They need to be in Power BI, and they need to be there now.” The business has urgent needs, and there’s pressure from management to see the result. But where do we start? What do we make? Defining the right requirements is essential if we want to make an effective report that is used and delivers value. Otherwise, it will just be collecting digital dust & cobwebs in an abandoned app… Here I will discuss 3 common approaches to collecting reporting requirements, arguing in favour of the third approach - working iteratively & collaboratively with business users.

The last few years have been difficult for most - if not all - of us, for different reasons. There seem to be more people struggling with stress & burnout than ever before. What concrete steps can we take to protect & improve our mental health? In this session, I introduce a 5-step process to help us overcome challenges like procrastination, imposter syndrome, anxiety and stress based on methods supported by mental health professionals [1]. The objective of this session is not to talk about myself or to speak vaguely, rather to zoom in on concrete tools & techniques to help overcome negative thoughts or experiences... to resist these 'gremlins' that hold us back. You can do this, dammit; I believe in you. More info: https://data-goblins.com/power-bi/mental-health [1] This talk has been audited by a professional psychologist to ensure integrity and avoid misinformation.

Due to the interactive nature of the session, participation is limited to 12 attendees. It is recommended participants have a basic technical grasp of Power BI and are comfortable doing groupwork with others. Feel free to also look at the visual agenda, available below: https://bit.ly/3B4vLPb -------------------------------------- ABSTRACT: -------------------------------------- A successful reporting project is about more than following and ticking boxes next to a requirements list. To design and launch a successful Power BI report, there must be sufficient understanding of the business users, their needs and how they use and think about the data. This is best done by co-creation - working with the users to design something that gets closer to the target. In practice, however, this is easier said than done. The objective of this session is to learn about user engagement and report design practices both from prepared content and each other. Attendees will work in groups to create and present a business case proposal and Power BI report design using sample data and user co-creation. This session will thus be an interactive, full-day workshop exploring how to engage with users and work hands-on with Power BI report design best practices. Attendees will be presented with a business case and sample data, then work in groups to engage with fictional user personas during interactive, role-played workshops. During the session, attendees will have to overcome typical reporting project challenges such as data availability, quality, unclear/changing requirements, technical or analytical challenges, etc. To overcome these challenges, attendees will have to work together and think outside the box, and engage with users to be able to design an effective report. At the end of the session, attendees will present the proposed solutions and compare to how they would have done it if they had only followed the requirements list. The format of the session will be styled similar to a tabletop roleplaying game (tRPG). Attendees will engage with a game master (GM), who will roleplay different stakeholders and users (yes there will be weird voices and accents), and narrate the events surrounding the fictional reporting project, using dice to introduce random obstacles and challenges that attendees will have to solve to succeed. PART 1: INTRODUCTION (1 H.) The first part of the session serves as an introduction for the rest of the day. We will look at the materials for the mock business case, and also experience an introduction to a helpful principle for data design and best practices. PART 2: VACUUM DESIGN PROPOSAL (0.5 H.) The second part involves some solo work. Using the requirements and provided materials for the business case, we sketch a proposed report solution from existing templates. The objective is to propose an effective design based on the information and scope provided. PART 3: CO-CREATION WORKSHOPS (5.5 H.) In user co-creation workshops, groups engage with users role-played by the organizer. The objective is to learn who the users are and what they do, in the context of the data. This is how they define the report, and together with the users, the design and functionality. PART 4: DESIGN SHOWCASE (1 H.) In the final part of the day, groups will present their designs and explain how user feedback was incorporated. We will discuss how the designs and proposals changed before vs. after engaging with users. Finally, we will recap the key take-aways from the session. -------------------------------------- IMPORTANT INFO BELOW: -------------------------------------- Are you submitting registration for this precon? If so, you can fill out this form, it will help us prepare! https://forms.office.com/r/NQm4sdRQmm

Data literacy is fundamental for an effective organization and is a keystone for successful data & analytics projects. It is clear, however, that data literacy is a challenge area in many businesses and society as a whole. A 2018 survey of business decision-makers found only 24% were “confident in their ability to read, work with, analyze and argue with data” [1], a finding recapitulated in the lay public [2] and demonstrated plainly during the global COVID-19 pandemic [3]. To traverse these rapids, we are in need of more than tools, but a wide knowledge of how they are used and the data concepts behind them. In this talk, I use a hypothetical business case to: I. Illustrate the concept of data literacy and the threat poor data literacy poses to an organization. II. Explore possible reasons why average data literacy is low and what we could do about it. III. Define practices and report design techniques to improve data literacy, such as points-of-engagement, data dictionaries, annotations, and more. Without data literacy, we will fail to navigate the waters of information that are ever-rising; we risk not only falling overboard, but even drowning. However instilling data in organizational values, engaging people with and about information in the right ways, and implementing accessible design practices can not only keep us afloat, but propel us onward to brave new worlds of effectiveness and value. [1] How to Drive Data Literacy in the Enterprise (2018) [2] Borner, K. et al. (2016) Investigating aspects of data visualization literacy using 20 information visualizations and 273 science museum visitors. Information Visualization 15(3) 198-213. [3] Advancing Data Literacy in the Post-Pandemic World (2021)