Conferences, user groups and community events.

Designing Production-Grade Custom Visuals in Power BI with Deneb

As Power BI practitioners mature, many eventually hit the same wall: native visuals are powerful, but they impose structural limits. Layering complex KPIs, building precise layouts, or controlling visual behaviour beyond the format pane often requires compromise.

This session explores how to move beyond those constraints by designing production-grade custom visuals using Deneb and the Vega-Lite grammar of graphics.

Rather than focusing on basic syntax, we will examine:

How to think in layers, marks, and encodings

When in-visual transformations outperform DAX measures

Structuring maintainable and reusable visual specifications

Handling dynamic scaling, conditional logic, and controlled interactivity

Avoiding common performance and maintainability pitfalls

Through real-world examples, we will rebuild complex reporting scenarios that native visuals struggle to represent cleanly — such as layered KPI cards, dynamic bullet charts, and composite visuals that would otherwise require multiple visuals stitched together.

This session is aimed at experienced report creators, data architects, and consultants who already understand Power BI modelling and DAX, and want to take control of the visual layer itself.

Attendees will leave with a clear framework for deciding when Deneb is appropriate, how to architect custom visuals responsibly, and how to elevate their reporting craft beyond default components.

Everyone’s talking about AI right now.

Boards are pushing for it. Leaders are being told to move fast or risk falling behind. So projects start. Budgets get signed off. Expectations go up.

But there’s something sitting underneath all of that that doesn’t get talked about enough.

The data.

Not because people don’t care about it… but because most of the time, they can’t actually see it.

That’s the gap.

The people making decisions about AI don’t have a clear view of the data those decisions depend on. And the people who do understand the data aren’t the ones setting direction.

So AI moves forward… while the biggest risk stays hidden.

And that risk is simple.

Bad data can cost you millions.

In some cases, a lot more.

The blocker isn’t AI. It’s the data.

And the problem isn’t just that the data has issues… it’s that those issues aren’t visible to the people making decisions.

Bad data isn’t new. But with AI, the consequences are bigger than ever.

AI needs strong foundations. Build on weak data, and you’re building on sand. It might look fine at first, but it won’t hold.

This isn’t about adding AI as a feature. It’s about business survival.

The organisations that get their data right will lead. The ones that don’t will fall behind.

And this isn’t something you can afford to wait on. This is about decisions you’re making now, not six months from now after a proof of concept tells you what you could have seen earlier.

Take a simple example.

Your sales data shows 1,000 products sold. 999 at £1. One at £100.

Looks fine.

Except it isn’t real. Someone missed a decimal point.

Now imagine building forecasts, pricing, or AI models on top of that. You’re not just slightly wrong… you’re confidently wrong. And AI will scale that mistake.

In this session, we bring both sides together. The technical view and the business view. Because this problem sits right in the middle.

Between us, we’ve seen this play out across organisations like Bank of America, the NHS, McDonald's, and UK policing. Different industries, same pattern. Data quality quietly shaping decisions at scale.

What we focus on is making that visible.

Using Microsoft Power BI, we bring together data from across your organisation. Finance spreadsheets, HR systems, platforms like Dynamics 365. One view of what’s really going on.

From there, you can see it clearly.

Where things don’t line up. Where risk sits. Where opportunity is being missed.

We look at data quality, governance, platform maturity, and operational risk, but more importantly, what those things actually mean for the business.

A key part of this is exposing what’s usually hidden.

Simple techniques like spotting outliers and inconsistencies quickly bring problems to the surface. The kind that would otherwise stay buried, but have a real impact on AI outcomes.

This takes AI readiness out of a dark room and puts it in front of the people making decisions.

Clear. Visual. Hard to ignore.

So instead of guessing, leadership can see where the organisation really stands, what needs to improve, and when it actually makes sense to invest in AI.

Power BI isn’t the focus.

Getting your data into the right shape is.

Power BI is simply the most effective way to make that visible, understandable, and something you can act on straight away.

If your data is ready, you move forward with confidence.

If it isn’t, you leave with something just as valuable. A clear roadmap of what’s in the way, and what to fix first.

Because AI doesn’t fail because the models are bad.

It fails because the data is.

Men in tech are great at shipping code, fixing problems, and keeping things moving. We’re much worse at talking about how we’re actually doing.

This panel brings together men from across the tech and Power Platform community to have an honest, practical conversation about mental health. Not theory. Not platitudes. Just lived experience.

We’ll talk about pressure, burnout, imposter syndrome, isolation, masculinity in the workplace, and what “coping” really looks like when deadlines, expectations, and life collide.

This isn’t about having all the answers. It’s about starting better conversations, normalising support, and reminding people that you don’t have to struggle in silence.

AI isn’t here to replace Power BI developers, it’s here to make us dangerously efficient. And while everyone is talking about Copilot, the reality is that ChatGPT, Claude, Grok, and even Gemini can massively accelerate modelling, DAX writing, documentation, design, debugging, and stakeholder communication.

This session is a hands-on, example-packed tour of AI workflows that Power BI people can use today. I’ll show the prompts I genuinely rely on: generating DAX variations, validating logic, rewriting gnarly measures, designing report layouts, creating SVG wireframes, summarising datasets, explaining model choices, generating test data, accelerating requirements workshops, and even turning stakeholder rambling into structured acceptance criteria.

This is not hype. This is not “AI will change everything.” This is the actual prompt pack that saves hours, improves quality, and makes you feel like you have a junior developer, a documentation assistant, and a design coach sitting next to you, without the HR paperwork.

Key Takeaways

Practical AI workflows to supercharge modelling, DAX, and design.

When to trust AI — and when not to.

The exact prompts you can steal and use immediately.

How to blend AI + Power BI + Fabric for faster delivery and clearer insights.

You've seen the demos. One model, one chart, everything works first time. Real builds don't go like that, and a room full of developers knows it.

So I built a complete three-page Power BI report through AI, end to end, without opening Desktop until the final check. This session is the honest account of how.

We'll walk the full pipeline across the three layers most AI content skips:

The model: tables, relationships, and measures, built by conversation through the Power BI modelling MCP. The design: the step nobody covers, wireframing the report in Figma first so it looks designed, not assembled. The report: pages, visuals, bindings, and theme generated through pbi-cli. Then the part that earns your trust: the twelve things that broke, and the fifteen-second fixes for each. You'll leave with the real workflow, not the demo.

Three takeaways:

The three-layer pipeline for building a Power BI report with AI, including the design layer almost nobody talks about. The specific gotchas that break a Desktop-free build, and how to fix each one fast. A clear sense of which parts of the pipeline are ready today and which still need your hands.

Every week there's a new AI tool promising to change how you build Power BI. Most won't survive contact with real work. I've run them against real models, real deadlines, and real client deliverables, and this session is the verdict.

No survey of the whole field. No hype. Just the tools I actually reach for, and exactly why they earn their place.

We'll climb three rungs, each one a live demo using Claude Code connected directly to a real Power BI model:

Chat-level AI, the ChatGPT and Claude tools you already use, and where they quietly fail you. Context-aware AI that reads your actual semantic model instead of guessing at it. Agentic AI that doesn't just answer questions but acts: writing measures, documenting models, and committing changes for you. You'll leave knowing which rung you're on today, what the next one looks like, and the one concrete step to climb it.

Three takeaways:

The difference between AI that guesses at your model and AI that reads it, and why that's the whole game. A live look at an agent writing, documenting, and version-controlling a semantic model. A curated shortlist of the tools worth your time, from someone who's discarded the rest.

“Can Someone Turn the Heating On?”

A manufacturing company had a problem.

Their internal testing failure rate sat comfortably below 1%. Their customers, however, were seeing replacement rates as high as 15%.

The strange part? It wasn’t consistent.

Some customers barely experienced issues at all. Others saw failures come in waves. Summer spikes. Winter spikes. Complaints would rise, disappear, then return again months later.

The data looked good. The engineers were smart. The dashboards were working. So why was the real-world experience telling a completely different story?

In this interactive session, I’ll walk through the accidental investigation that uncovered the issue, not through advanced AI, complex modelling, or some magical DAX formula, but through curiosity, context, and asking what felt like very dumb questions.

Questions like:

What’s different between testing and storage? What happens after the product leaves the workshop floor? Why are certain customers affected more than others? And eventually: Why is nobody putting the heating on?

This session is part Power BI story, part consultancy lesson, and part reminder that data professionals are not just report builders, we are storytellers. Sometimes the numbers only make sense when you step away from the dashboard and start understanding the humans, processes, and assumptions behind them.

You’ll leave with a practical framework of questions every consultant and analyst should ask, along with a different perspective on how to approach problem solving, stakeholder conversations, and the stories hidden inside your data.

Most importantly, you’ll leave asking better questions.

A live, end-to-end Power BI build

This is a demo-first session. No slide decks full of theory. No pre-built models. No “imagine if” examples.

The only slides you’ll see at the start are my face and a highlight reel of things I’m unreasonably proud of—purely for credibility, ego maintenance, and to prove I’ve broken Power BI in enough ways to know what actually matters. Then we switch straight into Power BI and build a real dashboard from scratch in 45 minutes.

We’ll start with raw, messy data and finish with something you’d genuinely feel comfortable putting in front of stakeholders.

Along the way, you’ll see:

How to decide what not to build (the fastest win in Power BI)

A practical approach to modelling without overengineering

The small number of measures that usually drive the most value

Layout and visual choices that prioritise clarity over decoration

How to keep momentum when time, scope, and attention are limited

This session is about decision-making under constraint—the same constraint most of us work under every day.

You’ll leave with a repeatable approach you can use when “Dashboard in a Day” simply isn’t realistic.

“Can you build us a management dashboard?”

Most Power BI projects start exactly like this, with a vague request and the expectation that the developer will somehow turn it into something insightful, beautiful, and easy to use.

Too often we jump straight into Power BI and start dragging visuals onto a canvas. The result is dashboards that look cluttered, inconsistent, and difficult to scale.

This session takes a different approach: design first, build later.

Starting with a typical stakeholder request, we will walk through the process of designing a Power BI dashboard from scratch using wireframes. Together we’ll decide what belongs on the page, how information should be prioritised, and how layout, colour, and typography influence the way users read data.

Along the way we’ll explore questions that good report designers ask before opening Power BI:

Why are we using these colours?

Is the design accessible and readable for all users?

How will this layout scale when more pages or reports are added?

How do we turn stakeholder requests into a structured visual plan?

By the end of the session we will have produced a complete dashboard wireframe and a repeatable design approach that can be applied to any Power BI reporting project.

What You’ll Learn

How to translate stakeholder requests into structured dashboard wireframes

How to design layouts that prioritise clarity and usability

How colour, typography, and spacing affect data interpretation

How to design reports that scale across multiple dashboards and teams

Key Takeaways

A repeatable wireframing process for Power BI report design

Practical design principles you can apply before opening Power BI

Techniques for creating dashboards that are clearer, more accessible, and easier to scale

Most dashboards are built backwards. This workshop is a deep dive into how dashboards should actually be designed in the real world — starting from messy stakeholder requests and working through to solutions a consultant would be proud of.

Most dashboards are built backwards.

They start with charts, colours, and whatever data happens to be available, and only afterwards ask what problem they were meant to solve. The result is something that looks impressive but fails the moment a real user tries to rely on it.

This workshop is not about where to click in Power BI.

It is about how to think.

This is a deep dive into how dashboards should actually be designed in the real world. Starting from messy, unclear stakeholder requests, we will break down how to turn vague requirements into clear outcomes, how to decide what matters (and what doesn't), and how to design reports that people actually trust and use.

Using real examples, we will explore common mistakes, poor design decisions, and the gap between what stakeholders ask for and what they actually need. From there, we move into practical application, working through a realistic scenario and designing a solution step by step, making the same decisions a consultant would make under real-world pressure.

This is a visual, discussion-led workshop designed for environments where attendees are focused on thinking, not following along on laptops.

If your dashboard needs explaining every time someone opens it, it has already failed.

This session is for anyone who builds, reviews, or relies on dashboards and wants to move beyond building reports into delivering real value.

Key Takeaways

• How to approach dashboard requests like a consultant, not just a builder
• How to identify what a dashboard should do before opening Power BI
• Common design mistakes and how to avoid them
• How to structure dashboards for clarity, usability, and trust
• A repeatable mental model for designing better reports

Session Format

• Half-day deep dive workshop
• Focused on thinking, not tooling
• Visual and discussion-led throughout
• Built around real-world scenarios and practical decision-making
• Includes live design exercises with audience input

A two-day immersive workshop transforming ethics, perception, and accessibility into a practical design framework for Power BI.

What happens when two philosophy enthusiasts who also happen to be Power BI experts team up? You get a two-day immersive workshop that transforms ethics,
perception, and accessibility into a practical design framework for Power BI.

Day 1 is dedicated to the theory behind responsible dashboard design. Inspired by Aristotle's persuasion pillars, Ethos (credibility), Logos (logic) and
Pathos (emotion), we explore ethical data design (Ethos), logical design aligned with human perception and cognition (Logos), and accessibility as an essential dimension of inclusive communication (Pathos). We examine how dashboards persuade, how visual hierarchy shapes decisions, and why accessibility is
not a feature but a responsibility.

Day 2 – We get practical. Inspired by John Dewey's philosophy, which challenges us to move beyond good intentions and focus on whether solutions genuinely
work when real people use them, this day is fully hands-on. Through live demonstrations and guided exercises, participants will storyboard their reports, capture assumptions and risks using data journals, prototype
layouts in Figma with AI-assisted exploration, and translate their designs into Power BI. Along the way, we embed accessibility validation, interaction
strategy, and usability testing into every step.

This is not another technical Power BI workshop. It is a design-first, ethics-aware approach to building dashboards that are persuasive, inclusive and
reliable.

A full-day workshop transforming ethics, accessibility, and human-centred thinking into a practical, repeatable design framework for Power BI.

What happens when two philosophy enthusiasts who also happen to be Power BI experts team up? You get a full-day workshop that transforms ethics, accessibility, and human-centred thinking into a practical, repeatable design framework for Power BI.

Every dashboard persuades. It shapes what people notice, how they interpret information, and the decisions they make. Over 2,000 years ago, Aristotle described persuasion through Ethos (credibility), Logos (logic), and Pathos (emotion), principles that remain surprisingly relevant to modern data visualisation and dashboard design. Complementing this, John Dewey's philosophy invites us to move beyond intent and focus on whether solutions genuinely
work when real people use them.

This workshop shows you how to translate intentional, ethical design principles into Power BI practice. You'll apply human-centred UI patterns and the WCAG
POUR framework (Perceivable, Operable, Understandable, Robust), then validate your ideas through wireframing and critique before building your report.