I've spent the last decade building Power BI dashboards for Australian businesses. I'm not going to do that work the same way for the next decade, because the market doesn't need more dashboards. It needs better decisions — and I can't in good conscience keep selling one when the real problem is the other.
So I'm relaunching my firm, Data Disruption, around a single idea that will either sound obvious or sound offensive, depending on how your last six dashboards went.
Here's what I mean.
Somewhere between 2015 and 2020, Australian mid-market businesses collectively spent hundreds of millions of dollars getting Power BI into production. It worked. Reports got prettier. Meetings got more visual. CFOs stopped asking for the pivot table and started asking for the dashboard. That was progress.
But sit in a mid-market finance meeting in 2026 and something uncomfortable is going on. The dashboards are beautiful. The executives nod. And then the business keeps making the same decisions it would have made if the dashboards didn't exist. The reports answer "what happened." Nobody ever asked them "what should we do."
I've watched this happen in logistics firms, in wholesale businesses, in professional services. I've built some of the dashboards in question. The problem isn't the data, and it isn't the people reading it. The problem is that dashboards were the end of the project — and nothing in the consulting industry's incentives pushes anyone to go further.
AI changes this. Not in the way the AI vendors would have you believe — not as a magic layer on top that makes everything better. It changes it because AI, properly applied, finally makes it economical to build the thing that should have existed all along: a semantic layer that answers the decision, not just displays the data.
I've been using Claude Code and the Fabric MCP in every engagement since late 2025. The delivery economics are genuinely different. Things that used to take six weeks take ten days. Things that used to cost $80,000 can credibly be priced at $22,000. And — this is the part that matters — the AI doesn't just compress the old work. It makes a new category of work affordable at mid-market prices. Custom AI agents grounded in a client's own semantic model. Forecasting embedded inside the reports people already read. Anomaly detection that drafts the board commentary before the analyst has opened the file.
I could have quietly raised my margins and told nobody. That's what most consultancies will do. Instead I'm writing this, because I think there's a better offer on the table for Australian mid-market businesses, and I'd rather bet on being useful than on being opaque.
So here's what Data Disruption does now, and — more importantly — what it doesn't.
- Three productised offers, at published prices. The Fabric Readiness Sprint is 7–10 days, fixed price. Ask Your Data is four to five weeks, one production-grade AI workflow built on your semantic model with a 200-question eval set. Predictive Power BI is three to six weeks per use case. All on the website. No "let's get on a call so I can quote you."
- No hourly billing. No T&M. Ever. Hourly billing punishes the thing that makes the new model work. Fixed scope, fixed price, or we don't take the work.
- No dashboards without a decision brief. If a client can't name the specific decision the report is meant to improve, we won't build the report. The discipline is the product.
- No black-box IP. On handover, you own the code, the semantic model, the prompts, the eval sets, the documentation. Claude Code wrote most of it; a human signed every line; and it's yours.
- One public artefact every week. I'll be writing here regularly — about Fabric, Copilot, AI agents, mid-market finance, and why your reports aren't changing decisions. No ebooks behind email walls. No "download the whitepaper." Just the work.
If this sounds like a manifesto, that's because it is. And if it sounds like I'm painting myself into a corner by publishing it, that's also correct. Writing it down publicly is the point. It's easier to refuse a bad T&M contract next month when the refusal is already on the record.
I want to be honest about who this firm is for, because a lot of consulting firms get vague about this deliberately. Data Disruption is for Australian mid-market businesses — roughly $50M to $500M revenue — with Power BI already in production and a board that wants an AI win this financial year. It is specifically not for ASX-50 enterprises (too slow), early-stage startups (wrong stage), or anyone who wants a 40-page strategy deck and a nine-month programme (wrong firm, politely).
A note on the name. I didn't pick "Data Disruption" to sound edgy. I picked it because after ten years, I think the category genuinely deserves disruption — not in the Silicon Valley sense, but in the mid-market Australian sense, which is quieter and more practical. It means building things that work on day ten, not on month six. It means publishing prices. It means refusing projects we can't deliver brilliantly. It means treating the client like an adult who can choose without a 90-minute discovery call.
I'll be writing here weekly. Some of it will be about Claude Code and Fabric and semantic models, because that's the craft. Most of it will be about the harder thing — the shape of the decisions Australian mid-market businesses are trying to make, and what it takes to actually help them make those decisions better, faster, and with more confidence.
If you run a finance function, a data team, or an operations group at a mid-market Australian business and any of this resonates, I'd like to hear from you. Not on a sales call. In a conversation. I've made a free 45-minute slot available for exactly this purpose — no deck, no pitch, no follow-up sequence. Tell me the decision you wish your data made easier, and I'll tell you honestly whether a Data Disruption sprint is what you actually need.
The website is datadisruption.com.au. The offers are public. The methodology is public. The manifesto is public.