Get people bought in
I want people on board with the goal and the process, not merely compliant with a decision.
Chris Barnes
My first AI-assisted project was a surrogate-key refactor across fifty-some ETL stored procedures, replacing on-the-fly identity columns with persistent keys so reprocessing a single fact table no longer required rebuilding the entire warehouse. It took an afternoon. Normally, it would have taken me a week. After twenty-something years running data warehouses and analytics functions at Comcast, Avaya, and other enterprises, including reporting for 45,000 customer service agents, I was amazed by that level of productivity acceleration.
20 years
10 in leadership
Running BI organizations at Comcast, Avaya, and other enterprises
45,000
Customer service agents on the BI reporting platform I owned
1 afternoon
vs. 1 week
My first AI-assisted project — a surrogate-key refactor across 50+ ETL stored procedures
Last year, AI models hit a quality threshold where their output didn't need constant second-guessing. I quickly pivoted to AI-assisted BI development. AI is an incredibly powerful tool, but it requires learning and experimentation. It wasn't an instant 10x accelerator. I kept running into bottlenecks, but I worked through them. Often, I would solve them on my own, then watch YouTube videos only to discover the official name for the bottleneck afterward. I was committed to learning AI-assisted development as fast as I could. I even bought a Mac Mini and spent over 100 hours building a personal AI assistant to better explore the technology.
AI is an incredibly powerful tool, but it requires learning and experimentation. Not everyone wants to watch AI YouTube videos in their spare time.
The frontier AI models are powerful and mature, but productivity and capability gains come only when you connect them to your data and business context. Right now, tech-savvy people like me are figuring this out, but these one-on-one AI projects are not scaling evenly across organizations. Not everyone wants to watch AI YouTube videos in their spare time. The next AI bottleneck is more of a people problem than a technology problem — the same kind of work I've been doing in BI for twenty years. That's what I want to work on next.
The next AI bottleneck is more of a people problem than a technology problem — the same kind of work I've been doing in BI for twenty years.
Why I got into BI
The best BI changes decisions, not dashboards.
My first real BI project was in 2006, on a small team launching a new credit company. The return-check rate was climbing, and we had only topline numbers, with no way to see which customers were driving it or why. Figuring that out wasn't the kind of technical work I'd been doing in network and system administration. It was closer to the business, and it changed how the business addressed the problem. It was the combination of business problems and technology that drew me into data and analytics, and that first project started my twenty-year career in business intelligence.
Why I'm so bullish on AI
AI changed what's worth building, not just how fast I work.
4 weeks
vs. 5.5 months
Third ERP added to our data warehouse. The first two had taken 5.5 months.
3 days
end-to-end
API connecting Salesforce to our warehouse. Sales people query real-time inventory while writing quotes.
6 hours
end-to-end
Custom BI planning tool. Replaced Excel; skipped the off-the-shelf alternatives.
The productivity gain is real. It took me five and a half months to add two ERPs to our data warehouse before AI. With AI-assisted development, I added the third in four weeks. But what gets me so excited about AI isn't just the productivity improvement. It's the capabilities the tool unlocks. I started attempting things I never would have before. In three days, I stood up an API that connects our Salesforce environment to the warehouse, so salespeople can query real-time inventory while writing quotes. That's the kind of cross-system project I would have written off as more work than it was worth. I built a custom planning tool for our BI team in six hours, replacing a clunky Excel process and skipping the off-the-shelf alternatives that always have too little or too much. Neither of those would have cleared the cost-benefit bar a year ago.
A year ago, the question was “Can I afford to build this?” Now it's “Is this actually the right thing to build?”
What hasn't changed is the process: requirements, profiling, architecture, design, validation, deployment, monitoring, and feedback. AI allows me to spend less time coding and analyzing data and more time working with the business, which is the most important part of my job.
How I Work
Useful change depends on collaboration, not just on better tools.
In medium and large organizations, a good technical idea only matters if people understand it, trust it, challenge it, and can adopt it. I bring a collaborative style to that work: clear enough to lead, open enough to learn, and practical about the constraints every organization must navigate.
Stay close to the business
I stay close to the people making the decisions, not just the data they look at.
Bring a point of view, stay open
I bring a point of view without pretending I have the full picture.
Outside Work
Curiosity is a through line for me.
I grew up in Colorado and still spend as much time as I can in the mountains, especially snowboarding in the winter and hiking or biking in the summer. Travel and photography are a big part of my life, too; I enjoy exploring places and the patterns they reveal. That curiosity also shows up in what I read, watch, and listen to.
Start a conversation
What I'm excited to do next.
With twenty years of experience in BI organizations (including ten in leadership roles) and a recent year working directly with AI, I'm excited for what's ahead. While the tools may be new, my goal remains the same. If you're tackling similar challenges, I'd love to talk.
20 years
Building BI in enterprise organizations
10 years
Leading BI teams
1 year
Hands-on with AI as a developer