Clean the data myself first
I go through the raw dataset by hand before anything else touches it; missing values, duplicates, inconsistent types, outliers. This is where I actually learn and understand the data.
How I Work
I use AI to work faster. I don't use it to think for me. Here's exactly where AI fits into my process, and where it doesn't, because if you're trusting me with your business data, you should know the difference.
Use AI to catch what I missed
Once I've cleaned it, I run a second pass with AI to flag anything I overlooked. Every flag gets resolved one of two ways: it's a real miss and I fix it, or it's a deliberate choice for a business reason, and I document which.
Delegate the mechanical work, with review
Formulas, DAX measures, and first-draft write-ups are rule-based once the data and the goal are defined; AI is fast and reliable here. I verify every formula does what I intended before it goes anywhere near a deliverable.
Keep the judgment calls mine
What to highlight on a dashboard, which insight matters most, how to frame a recommendation for your business, that's not something I outsource. AI can draft; it doesn't decide.
Verify before delivery
Every number, chart, and formula is spot-checked against the source data before submission. Nothing ships that I can't personally explain and defend.
"If I can't explain how a result was produced without re-reading an AI's explanation, it doesn't go in your report. That's the standard on every project."
Why this matters to you
You're not getting a black box. You're getting analysis where a human being has verified every figure, made every judgment call, and can walk you through the reasoning behind every recommendation, with AI used to move faster, never to skip the thinking.