Raw Data with Rob Collie

Pivot tables finally auto-refresh. Humanity wins, but our host Rob Collie is still annoyed. Why? Because he asked for it back in 2007 and got shot down. That little gripe kicks off a bigger conversation: what else in data is overdue for a shake-up? Copilot might be next. Forget filters. Forget dashboards. Rob put it to the test on his beer-league hockey stats and found himself asking questions faster than any report could keep up. It felt natural, almost too natural. Then came the twist. Copilot served up an answer that looked perfect… and was completely wrong. If it can fool the guy who built the model, what chance does anyone else have? Listen in for the laughs, the lessons, and the curveballs of a future where filters may finally be obsolete.

Every campaign has it: that shiny "more reach, lower cost" lever that looks like marketing gold but really just siphons your budget into digital quicksand. We pulled it. The metrics looked fantastic. And that's exactly how we knew we'd been had. In this episode, Rob breaks down a real-world lesson in false signals, phantom clicks, and why data discipline isn't just consultant-speak—it's your financial survival strategy. We're talking about the bots that sneak in through "partner sites," how they corrupt your retargeting, and the ripple effect that turns good data bad across your entire funnel. Here's the thing: The real damage isn't the wasted spend. It's what corrupted data does to your decision-making. When you're building your strategy on phantom signals, every "optimization" takes you further from real results. If you're protecting a budget or leading a team that depends on clean data to drive real business decisions, this episode cuts straight to what matters: spotting the traps before they drain your resources and rebuilding trust in the numbers that actually count.

AI is rewriting the rules of analytics. Copilot can pull answers straight from your semantic model and bypass the dashboard entirely. But for all the tech fireworks, the same old truth holds: communication is still the hardest part. Stakeholders don't always know what they want, builders don't always know how to translate it, and requirements docs have never fixed that gap. Copilot just puts the tension in sharper focus. Rob and Justin dig into why vanishing chat histories aren't just inconvenient, they erase the most honest record of what stakeholders actually care about. Screenshots and Word docs are a band-aid, not a solution. Persistent, shareable conversations could change the way model developers and business users collaborate, but only if governance and security evolve fast enough to keep up. Along the way, they show why usage data from Copilot queries is miles ahead of click stats on a dashboard and why the story of your data has always hinged on the same thing: people understanding each other. Dashboards may have set the stage, but conversation is where the real action is. Listen now and see what happens when the chat itself becomes the deliverable.

Two hundred episodes in, and we're done with the warm-up! Episode 200 finds Rob flying solo and pulling zero punches on the question everyone's quietly asking: what's the future of data work? No anniversary nostalgia here, just uncomfortable truths about AI bootcamps at Starbucks, semantic models going naked, and why being "pretty good" at anything is about to get very complicated. If you think you know where this is all headed, think again. Rob spent yesterday turning a non-techie real estate agent into an AI power user, and what he saw exceeded anything from the early Power BI days. But here's what nobody's talking about: the invisible barriers, the shifting skill requirements, and why the middle ground might be disappearing faster than anyone realizes. The lines between data work and software work are blurring, structured versus unstructured data is becoming meaningless, and the comfortable assumptions about who does what are about to get stress-tested. Two hundred episodes of calling it straight, and this one tackles the questions that keep data professionals up at night. Some answers might surprise you. Others might make you uncomfortable. [But you'll know exactly where you stand when the dust settles]

Most of us have been in the trenches long enough to know when something's about to flip the script. And brother, we're standing at the edge of a cliff most data folks don't even see coming. Rob Collie thought he had Power BI figured out. Then Copilot did something impossible; it cracked a question that should've left it scratching its digital head. But it didn't just answer. It nailed it. That's what we're calling the Dobie Moment—when AI stops being a fancy calculator and starts being genuinely scary-smart. Here's the thing nobody's talking about, your semantic models aren't just sitting there anymore. They're waking up. And when Rob and Justin break down what happened in this episode, you'll see exactly why that should make you sweat a little. They're not here to blow smoke. They'll show you the magic, sure, but more importantly, they'll show you where the landmines are buried. Because when AI starts connecting dots you didn't even know existed, confidence and correctness become two very different animals. Bottom line: The future of data just knocked on your door. You can pretend you didn't hear it, or you can listen to this episode and actually be ready when your models have their own moment of reckoning. Your call. But don't say we didn't warn you.

AI looks unstoppable… until you hand it a hundred pages of meeting notes. Rob and Justin dig into why context windows and token limits quietly run the show. That "million-token" brag from Google? More like weighing the Titanic in bananas. From Shakespeare to SharePoint, this episode shows why AI remembers the Roman Empire better than your company history—and why that's not a bad thing. Rob also introduces Griff, a digital colleague that fires off P3-flavored ideas like it's had three espressos. It's practical AI that's actually fun to use. Hit play to find out where AI is brilliant, where it falls flat, and how to make it work for you without the hype. Also on this episode: Million Token Context Windows? Myth Busted—Limits & Fixes

Back in 2010, Tableau beat smarter tools with a better demo. No brain, all charm and the market loved it. Fast-forward to now: same playbook, new costume. The AI dashboard crowd is selling "natural language BI" with zero semantic model, zero memory, and a whole lot of LinkedIn swagger. In this episode, Rob and Justin revisit why Tableau's empty-calorie approach won the first round, and how that same mistake is about to flood the AI + BI space all over again. Turns out, you can still sell snake oil if you call it GenAI. Rob breaks down how an elite MIT course managed to skip LLMs entirely, how a flashy Tableau blog post went viral for connecting a CSV, and why "AI-ready" vendors keep duct-taping chat interfaces onto raw SQL and hoping no one looks under the hood. But the real story? Microsoft is sitting on the most powerful data brain in the game, and if they land the front end, it's game over. This isn't just a history lesson. It's a blueprint for seeing through the hype and betting on what actually works. If you're building, buying, or betting on AI tools, listen in before you get dazzled by the demo. Also on this episode: Early Experiments in Tableau's New MCP Service

Let's say your business runs on sun, sweat, and schedule precision. You've got crews in the field, materials that don't wait, and about 90 minutes to get it right before the product turns into a thousand-pound paperweight. That's the world Joseph Graziano lives in. He's the CFO of Quadrant Concrete and also the guy keeping the trucks moving, the forecasts dialed in, and the safety records spotless. Because in the mid-market, you don't get extra people. You get extra resourceful. Joseph shares how he helped transform a boots-on-the-ground concrete business into a data-forward operation without fancy titles, inflated budgets, or a fleet of consultants. From field-collected data to real-time cashflow forecasting, he's found the sweet spot where better reporting leads to smarter, calmer decisions. You'll also hear why operational transparency isn't just about ROI, it's about reducing chaos, building trust, and creating a culture where everyone sees what matters. If you're trying to lead your business through complexity without adding complexity, this episode can be your blueprint. Data doesn't have to be fancy to be powerful. It just has to work. Listen now and see what it looks like when grit meets insight.

It started as a side project. Rob Collie built a Power BI model for his rec league hockey team. Just for fun. Just to see what the data could say. But something weird happened. The dashboards were solid. The data model was solid. But, the users still had questions. And lots of them. And that's when it clicked: people don't think in slicers. They think in questions. Natural ones. The kind dashboards rarely anticipate. In this episode, Rob and Justin Mannhardt didn't just talk about Microsoft's Copilot for Power BI. They put it to the test. No tuning. No prep. Just a raw semantic model paired with real questions from actual humans. The result? A glimpse at what happens when the tech finally meets the moment. Copilot isn't just a gimmick. It understands nuance, handles filters, and points people to the answer without making them dig. And it's getting better by the day. This isn't a future-state conversation. You've already done the hard part. Now you can build on it. And if you've been wondering when AI will start delivering real value, this is a pretty good place to start. Also in this episode: Indy Inline Hockey Dashboards Inline Analytics Doesn't Mean What You Suspect it Means, w/Ryan Spahr Copilot for Power BI Rethinking the ROI of Dashboards

AI wants your data, but it can't handle the truth . . . "We'll just plug our data into ChatGPT." Sure. Sounds easy. Rob thought so too, right up until he hit token limits, had to install Python, and discovered that even two megabytes of transcripts was too much for the world's smartest models to handle. In this episode, Rob and Justin break down what happens when you try to use AI on your company's internal data. Spoiler: it's a lot more complicated than the vendors make it sound. From semantic search to retrieval augmented generation (RAG), they unpack why the dream of "ask AI anything about your business" keeps falling short. You'll hear why your Power BI model isn't going anywhere, why structured data still needs old-school engines, and what it really takes to get value from your own information. Somewhere between a cautionary tale and a tech detective story, this one tells it like it is . . . unapologetically. Run into the same situation? We'd love to hear about it! Give us a shout on LinkedIn and tell us how you overcame the limitations.