Tim Heng

Description: 

Danielle Stein Fairhurst, Ian Bennett, Liam Bastick and Tim Heng 

Join leading experts for a practical panel discussion on financial modelling. The session will explore current trends, common challenges, best practices and the evolving role of technology in building reliable financial models.

Kris Jaganah

Discover the similarities between Python and Excel, including formulas, functions, data analysis and automation. Learn how familiar Excel concepts can make Python easier to understand and use.

Ilgar Zarbaliyev

In this session, I will explain recursion in DAX and how recursive business logic can be handled in Power BI. We will look at how calculations can depend on previous results, why traditional recursion is difficult in DAX and how user-defined functions can help create more practical solutions.

Sam Ngo

Part 1 — The building blocks
Vocabulary for data shapes: scalar → vector → matrix → tensors. The key concept that “”rank”” is simply how many labels point at one number, and each dimension is just another axis.

Part 2 — Laying out n dimensions
Where each extra dimension physically lives in a workbook, demonstrated by growing one workbook step by step: flat grid → worksheets as a third axis (with bookend sheets) → sheet blocks as a fourth axis. It ends with the “”five ways to add a dimension””: label columns, sections, sheets, blocks and files, in any order.

Part 3 — Summarising with two engines
The payoff, and the longest section:

Array arithmetic fundamentals: broadcasting, element-wise operations, spilling grids, and the difference between * and MMULT
MMULT explained: the dot product, totals with a block of 1s, the “”collapse ladder”” (3D → 2D → 1D → 0D), the dimension dial, and matrix inverse
A worked example (9 records × 10 periods) showing MMULT as a “”switchboard”” and as a 2D SUMIFS
The “”pocket array”” technique in five steps, including SUBTOTAL and pockets inside SUMIFS

Closing
A neat hook connecting it all to AI (a neural network layer is just one MMULT), then three takeaways.

Jean Philippe Christian Mbolanantenaina

Laura Szepesi

“Modern Excel Dashboards: From Messy to Magical

Tired of boring, clunky spreadsheets nobody wants to open? In this session, you’ll see magic tricks that turn messy, old-school dashboards into stunning, jaw-dropping ones – the kind that make people stop scrolling and ask “”wait, that’s Excel?!””

You’ll walk away with:
→ Creative Excel workarounds that turn basic dashboards into polished, modern ones
→ How to level-up your slicers into beautiful, intuitive filters
→ Clever fixes for errors – the kind your users will actually thank you for

All that without VBA — just clever Modern Excel.

Whether you’re building reports for your team or a dashboard to impress your boss or clients, this session will give you the tricks to make it happen. Come for the WOW moments, leave with dashboards people actually want to look at.”

Olufemi Olamoyegun

Many reporting processes begin in Excel but become difficult to maintain as data grows. This session shows how to identify common reporting habits that slow analysis and how Power BI solves them with a modern reporting approach.

Through live demonstrations, we’ll move an Excel report into Power BI and improve it step by step.

Topics include:

• Importing Excel data into Power BI

• Building a simple star schema

• Replacing manual formulas with reusable measures

• Creating interactive reports with slicers and drill-through

• Publishing and sharing reports without sending multiple Excel files

Attendees will leave with practical techniques they can use whether they currently work in Excel, Power BI, or both.

Jan Karel Pieterse

Many Excel users have written a macro, created a UserForm or even turned a workbook into an add-in. Far fewer have released something to the public and supported it for years.

In this session, Jan Karel Pieterse shares lessons learned from developing and maintaining Excel tools such as Name Manager, RefTreeAnalyser, FlexFind, Spill Manager and Workbook Info.

We’ll look at what happens after version 1.0 is released: deciding which features to build, keeping users happy, surviving Excel updates, dealing with bug reports, supporting multiple versions of Excel, and making sure the tool also works on international versions of Excel and supports multiple languages.

If you’ve ever considered turning an Excel solution into a product, this session offers practical advice, hard-earned lessons and a few war stories.

Frédéric LE GUEN

We all produce documentation, specs and reports in Word — and most of them have no real structure. Every heading is “”just bold, 14pt””, so there is no navigation pane, no automatic table of contents, no reliable PDF export, no accessibility, and no way for the next person to edit the file without breaking it. Cleaning that up by hand, paragraph by paragraph, is the kind of task nobody ever finds time for.

This session shows how I solved it with StyleCleaner, a Word add-in I built after years of doing it manually.

What we will cover:

– Why unstructured Word documents cost more than people think, and why manual clean-up doesn’t scale
– A full demo on a real, messy document, through the add-in’s three tabs: Analysis, Designer, Cleanup
– How the detection works: font size, bold, alignment and indentation are read to decide whether a paragraph is a Heading 1 to 4, body text or a list item
– What you get back: a working navigation pane, an automatic table of contents, a consistent look, and a document that survives being edited by someone else
– Where the tool stops, and how a rules-based approach compares with asking generative AI to reformat a document for you
– Questions, plus a discount code for EVG attendees.

Ana María Bisbé York

Data transformation and cleansing involve multiple challenges. Typically, data analysts draw on data from external sources that they don’t own, so they adapt to what they receive, when they receive it, and in what state.
These challenges must be overcome before moving on to the next stage of data processing and visualization.
Power Query is a data preparation and transformation engine. It includes a graphical interface for retrieving data from sources and a Power Query editor for applying transformations using the M language.

Charles Williams

The session will cover:
– Finding the slowest formulas in your workbook.
– What makes some formulas slow
– Making slow formulas faster
– Are dynamic array formulas fast or slow.
– Different ways of using lists in FILTER and SUMIFS etc
– Examples of speeding up slow formulas.

Kathryn Newitt

Can you create sophisticated Power Query transformations without writing the M code yourself? In this session, I compare the intelligence already built into Power Query with using generative AI to create queries from natural-language prompts. Through two demonstrations, you will see the strengths, limitations and risks of each approach, and learn how to turn AI-assisted code into a reliable, refreshable solution.

Coming Soon!

Mike Thomas

PivotTables have been Excel’s go-to tool for summarising and reporting data for decades. But modern Excel now offers another way, one that gives you far greater control over both your calculations and the appearance of your reports.

In this session, we’ll move beyond the PivotTable and use modern Excel formulas to create flexible, professional-looking reports that update automatically as your data changes.

You’ll discover how Dynamic Array Functions such as FILTER, SORT, UNIQUE and GROUPBY can handle many of the tasks traditionally performed by PivotTables, while allowing you to design the layout exactly as you want it.

Bill Jelen

Use ChatGPT or Copilot to automate the most boring parts of your job! Replacing eighteen minutes of monotony with a custom GPT in ChatGPT or custom Skill in Excel Copilot is easy and requires no code.

Jon Peltier

Celia Alves

Celia will introduce the recently added features of Lists, Arrays in Cells, and Nested Arrays, by looking at how they work and a few simple application examples.

Sam Ngo

Part 1 – Meaning as vectors

How a word becomes a row of numbers (an “embedding”). What D1…D50 really are (fifty scores the computer invents), and how GloVe learns them from which words keep the same company. Similarity is a dot product (SUMPRODUCT); with every word scaled to length one, it becomes the cosine between two arrows. Then:

– One MMULT gives a whole similarity grid, and one MMULT over 10,000 words is vector search (Sydney → Melbourne, Adelaide, Brisbane)

– A 2-D map of meaning, and a spelling fix for typos using three-letter pieces and a Jaccard score

– Why this is how “chat with your documents” (RAG) works

– Word arithmetic: king − man + woman ≈ queen, paris − france + italy ≈ rome

Part 2 -Predicting the next word

Shannon’s 1948 idea: count what comes next. The bigram table is one MMULT of two 0/1 switch matrices, and dividing by row totals turns counts into probabilities: a tiny language model. Each extra word of context adds one more dimension (vector → matrix → cube → tensor) until the tables explode. Almost every cell is empty, so the trigram cube is stored flat with GROUPBY: know when to go flat.

Part 3 – Attention and the transformer

Why a word needs its context (bank next to river vs money), then attention built step by step on “the king and his queen”: word vectors plus position waves, then Q, K, V as three MMULTs, scores, the no-peeking mask, softmax, and the blend (queen borrows 23% from king). The transformer block in plain words: look around, keep and tidy, think, repeat, pick the next word. It ends by comparing the same machine at GPT-3 scale.

Closing

Why AI is really “all MMULT” (over 95% of the maths, and what GPUs are built for), what training adds, then three takeaways: meaning is a vector, prediction is counting or mixing, and layout still decides everything.

Talia Cao

Top N seems straightforward until you start dealing with ties. Then the question becomes: what exactly do we mean by the “Top N”?
In this session, we’ll go beyond the standard Top N in Power BI and build a set of DAX measures for different ranking scenarios:
1. Top exact N values
2. Top N including ties
3. Top unique N including ties, with dense rankings
4. Top unique N including ties, with skipped rankings
5. Top unique N excluding ties.
We’ll break down the DAX logic step by step and look at how each approach handles duplicate values and rankings.
This session is aimed at Power BI users with a basic knowledge of DAX, especially finance and accounting professionals who want more control over Top N reporting and rankings.

James Tomas

Learn how to build flexible Top N rankings in Excel and take control of how ties and duplicate values are treated. This session covers several ranking approaches and the formulas behind them, giving you more control over how your results are displayed and analysed.

Ian Bennett

Coming Soon!

Nabil Mourad

In this tutorial I show you how to prompt AI automatically (Claude or Copilot) every time your data changes and get insights and recommendations about a work situation without the need of opening a Claude or Copilot Side pane or retyping a prompt. Insights are always delivered to the worksheet with formula recalculation.

Indira Bandari

Ilgar Zarbaliyev

In this session, I will create a declarative agent using generative AI, refine its instructions, publish it to Microsoft 365 and Microsoft Teams and test it in Microsoft Copilot.

Duncan Williamson

In this practical session, Duncan Williamson puts AI-generated Excel work under the same scrutiny that an experienced modeller would apply to a human built workbook. Using real modelling experiments, he will examine dynamic array and LET formulas, investment appraisal and scenario modelling, Monte Carlo simulation, and examples where apparently convincing results concealed errors in assumptions or workbook construction.

Rather than presenting AI as either a miracle or a menace, the session develops a practical approach to working with it: Specify → Generate → Test → Challenge → Reconcile → Improve.

The objective is not to replace spreadsheet expertise with AI, but to discover how AI can make an experienced Excel user more productive: provided we continue to question the answers.

Craig Hatmaker

Dynamic arrays make models much faster and inconsistent formula errors impossible. But to model with Dynamic Arrays requires advanced skills – unless we use 5g functions.
5g functions are easier to use than the formulas they replace. They completely eliminate the technical barrier to modeling with dynamic arrays and, in the process, eliminate errors.
In this session, we will show common financial modeling errors, some of which caused catastrophic losses, that dynamic arrays and 5g functions make impossible.
We will also demonstrate how to create your own 5g functions by using Copilot to convert a traditional model section to a 5g function.
We will use that function to benchmark dynamic array speed compared to traditional formulas.
We will then show a traditional 3 statement model and compare it to the same model converted to dynamic arrays with 5g functions.
Lastly, we will show how to get these functions for free.
Faster performance and fewer errors made easy – for free.

Danielle Stein Fairhurst

Almost every financial model has a time-based element and building a flexible timeline is the cornerstone to beginning its construction. It’s important to get the time frames right front the outset, and if you don’t, subsequent adjustments can waste time and introduce errors. By using Excel’s dynamic array functions to build the timelines, we can adjust both the unit of time measurement and the timeframe of the model instantly, making the model robust, reliable and easy to use. This not only enhances usability but also expedites the model-building process. In addition, the use of a variable start date, when applied correctly, enhances the model’s durability and adaptability over time.

Gildas Raharimanana

Comparing two versions of an Excel workbook is a common requirement when reviewing financial models, but manually identifying changes can be difficult and unreliable. This presentation will demonstrate how Microsoft Excel’s Spreadsheet Compare and Inquire Add-in can be used to systematically investigate changes between two versions of a workbook.

The presentation will cover:

Why comparing different versions of a workbook is important in financial modelling and model auditing
How to use Spreadsheet Compare to identify changes in values, formulas, formatting, worksheets and other workbook elements
How the Inquire Add-in can help analyse workbook structure, relationships and dependencies
A practical demonstration using two versions of a financial model containing deliberate changes
How to distinguish between changes that are merely cosmetic and changes that can have a material impact on the model
The key limitations of Spreadsheet Compare and Inquire, and why these tools should support rather than replace professional model review
The presentation will ultimately show how these tools can help answer two important questions: “What changed?” and “What could that change affect?”

Henk Vlootman

In this session, I’ll share how Microsoft Fabric and Power BI helps to bridge the gap between analytics and business operations. We’ll explore how organizations can move from insights to action in a single platform, using real-time data, automation, and integrated analytics to drive faster, smarter decisions. W’ll show how to use the new Power BI slicers and how it works within this proces.

Melissa de Korte

Sometimes the most enjoyable way to learn is to give yourself a problem and see if you can crack it.

In this session, we’ll tackle several Power Query challenges, examine the data, identify patterns, and build solutions in M. The goal isn’t just to arrive at the answer, but to strengthen the problem-solving instincts you can take back to completely different problems.

Miguel Félix

How many times do users ask you: can I export to Excel?

Our first reaction after building lakehouses, warehouses, semantic models, and other modern data environments is often to resist. But what if we start looking at this request with fresh eyes?

In this session, we explore how Microsoft Fabric integrates seamlessly with Excel. Move beyond simple data dumps and discover how to establish live connections to semantic models, leverage SQL endpoints on lakehouses and warehouses, and streamline data exploration directly from notebooks and pipeline copy activities.

Learn how Excel functions as a natural extension for consuming and feeding data into OneLake rather than a tool to run away from. Join us to see how Excel isn’t being replaced—it’s being supercharged.

Live Connections: Connect Excel directly to Fabric semantic models for enterprise-grade reporting without data duplication.

SQL Endpoints: Query Lakehouse and Warehouse endpoints natively to bridge traditional workflows with cloud data stores.

Modern Workflows: Export data explorations smoothly from notebooks and pipeline copy activities into familiar spreadsheet interfaces.

Cristian Angyal

Markus Ehrenmueller-Jensen, Olufemi Olamoyegun, Henk Vlootman and Ilgar Zarbaliyev
Join our panel of experts for a practical discussion on Microsoft Fabric and Power BI, exploring current trends, real-world applications, integration and the future of data analytics.

Markus Ehrenmueller-Jensen

Key-value-pair tables – often used in applications for their flexibility – typically consist of just two columns: a key (or attribute name) and a value. This format allows developers to easily add new fields without altering the table structure, making it a favorite in application development.

But what’s great for developers can be a major headache for report builders. Business Intelligence tools like Power BI thrive on structured, columnar data – where each attribute has its own dedicated column and correct data type. To get there, key-value tables must be pivoted and typed properly – a task that can be tedious, error-prone, and difficult to maintain.

In this session, Markus Ehrenmueller-Jensen shares a real-world solution he implemented for a client running in production today. He’ll walk you through a semi-automated approach for transforming key-value tables into analysis-ready datasets—both in self-service BI scenarios using Power Query, and in enterprise-grade pipelines using SQL and PySpark in Microsoft Fabric.

What You’ll Learn:
* Challenges of working with key-value data in reporting tools
* How to pivot and reshape key-value tables for analytics
* Automatically assigning correct data types for aggregated reporting
* Self-service vs. enterprise approaches in Power Query, SQL, and PySpark

Who Should Attend:
Power BI developers, data engineers, and architects working in Microsoft Fabric who need to turn flexible data structures into performance-optimized models for reporting and analysis.

Yolanda Cuesta Altieri

Business risks rarely emerge without warning; they often leave subtle clues in the data. This session Learn how to combine Excel spreadsheet, Copilot and financial expertise to shift from reporting past results to anticipating future risks.
We will see how to find hidden trends, flag anomalies, and identify emerging risks before they affect performance.
Topics:
• Detect early signs of margin erosion
• Assess customer concentration risk
• Recognize cash flow warning signs
• Uncover anomalies in financial data
• Turn AI-generated insights into stronger business decisions.

Rahim Zulfiqar Ali

This presentation provides a practical, hands-on guide to designing and building interactive descriptive analytical dashboards using Microsoft Power BI Desktop. Participants will learn how to connect and transform data with Power Query, create an optimized data model, develop DAX measures for key business metrics, and design visually appealing dashboards using charts, KPIs, slicers, and custom visuals.

By the end of the session, attendees will be able to create professional, user-friendly Power BI dashboards that deliver meaningful insights and support effective business decision-making.

Susana Pinto

In this 50-minute recorded session, we follow FlowBake – a fictional Portuguese bakery chain – to show how any small business can turn Excel into a fully functional AI-powered analyst, without Copilot, without Power BI, and without an IT department.

Using Power Query to automatically consolidate monthly sales files, and a VBA macro connected directly to the ChatGPT API, we build a live pipeline that analyses 330 real business data points across three departments – Sales, Marketing, and Customer Satisfaction. Each dataset generates AI-powered insights directly inside Excel cells, and the results feed into a dashboard that produces KPIs and a CEO Executive Summary crossing all three departments.

Every tool used is already available to any Excel user: Power Query, VBA, and an OpenAI API key that costs less than a coffee per month. Attendees leave with a downloadable template and the confidence to adapt it to their own business data – whatever sector, whatever system, whatever size.

Ester Bergen & Preethika Kiruveedula

Coming Soon!

Markus Ehrenmueller-Jensen

Do your visuals take forever to respond to filters?
Is DAX starting to feel more like a puzzle than a language?
Does your model view look like a tangled web of chaos?

Chances are, your data model is missing one crucial ingredient: a star schema.

In Power BI, a well-designed star schema isn’t just a best practice – it’s the key to performance, simplicity, and scalable DAX. The good news? Transforming a messy model into a clean star schema is easier than you think.

In this practical session, Markus Ehrenmueller-Jensen will walk you through common modeling pitfalls and show step-by-step how to reshape your model into a star schema using Power Query and DAX. You’ll learn how to:

* Assign attributes to fact and dimension tables
* Normalize your fact table and denormalize dimensions
* Add and configure a proper date table

By the end, you’ll not only understand why star schemas matter—you’ll be well on your way to becoming a Guardian of the Star Schema.

Who Should Attend:
Anyone working with Power BI who wants faster reports, simpler DAX, and cleaner data models. Ideal for beginners and intermediate users ready to level up their modeling game.

Ken Puls, Bill Jelen & Aditya Kumar Darak

Join our expert panel to explore how AI is transforming Excel, from smarter data analysis and automation to practical real-world applications and future opportunities.

Bernard Boateng

AI can now build you a dashboard from a single prompt. It suggests formulas, formats charts, and structures your data before you’ve touched a PivotTable. So where does that leave the Excel professional?

In this session, we will put today’s AI tools to the test live: building a real dashboard with Claude/Copilot to see exactly what AI gets right and where it still needs a human hand.

Walk away with a clear, practical answer to the question every Excel user is asking: what should I let AI do, and what should I still do myself?

No hype. No fear-mongering. Just an honest look at where the line sits, and how to work on the right side of it.

Echo Swinford

The best charts communicate their message almost instantly. This session focuses on the design side of data visualization – not which chart type to choose, but how to make the charts you create easier to understand. We’ll explore how our brains process visual information and how to apply those principles when designing charts. You’ll learn practical techniques for emphasizing what’s important, removing distractions, improving labels and titles, and preparing Excel charts that look polished both in Excel and when presented in PowerPoint.

Cristiano Galvao

Ilgar Zarbaliyev

In this session, I will show how to create a Power Apps canvas app using Copilot. We will look at how Copilot can build an app from an Excel file or from a simple description of the data, using a Dataverse-enabled environment.

David Benaim

Many dismiss Google Sheets as inferior, but actually Google does many things better than Excel (and not just collaboration). In this demo heavy session, David will show features such as Google’s superior data entry features and Tables, functions in one but not the other and how AI tools vary across the platforms, in particular two useful features that Google Gemini in Sheets has which don’t exist in Excel, Claude or Chat GPT add ins.

Ilgar Zarbaliyev

In this session, I will show how to create a scheduled Power Automate flow using the recurrence trigger. The flow will read email addresses from an Excel workbook in OneDrive and automatically send event information once a day.

Aditya Kumar Darak

– Create your first agent in minutes (no code required)
– Build a smarter agent that reasons through reconciliations and variance queries
– Connect your agent to your data sources: Excel, SharePoint and your accounting system
– Build a workflow agent that captures invoice details on arrival and routes them for approval
– Share your agents in Teams and track how they are used.

Michael Olafusi

Liam Bastick

Having worked with Excel since its very inception, Liam reflects on how Excel has evolved over the years, where it is now, and what the next few years might hold given AI, BI, dynamic arrays, nested arrays, LETs, LAMBDAs and of course, the mighty FACTDOUBLE.