start here
Unprompted
AI agents are starting to act before we ask. Drafting, deciding, creating on our behalf without us ever naming the goal. What happens to human agency when the friction of being asked disappears?
Derek Özkal · Kansas City, Missouri
Twenty plus years across economic research, philanthropy, and journalism, most of them at the Ewing Marion Kauffman Foundation, where I've directed $31.9 million in research on entrepreneurship, labor markets, and what AI is doing to both. Agents and Agency is the place where you can find my work about keeping human dignity, judgment, and meaning at the center of the AI debate. This site is the paper trail. Or digital trail, anyway.
$31.9M directed · 125+ grants · 90+ convenings · 35+ bylines
Agents and Agency is where I write now: research-based essays on AI, work, and human agency, with too many footnotes (but that's where the jokes live!). Below you can find links to my current writing, Kauffman Currents pieces, Kansas City Business Journal articles, and other work.
start here
AI agents are starting to act before we ask. Drafting, deciding, creating on our behalf without us ever naming the goal. What happens to human agency when the friction of being asked disappears?
New essays land on Substack first. Written by me, edited with AI, argued with until it sounds like me. No noise between issues.
I have had the opportunity to work with so many great researchers. The job is to fund the right questions, read the drafts, make relevant introductions, and get the findings to people who decide things in a format that they can actually use.
Through the Kauffman Knowledge Challenges and related programs: $12M deployed across 60+ projects, selected from 1,100+ submissions.
A few of the papers, articles, or reports that were kind enough to thank me, quote me, or reference my work.
Invited philanthropy participant in the 2021 NSF workshop (regional and place-based innovation track) that NSF pointed to as part of the groundwork for its Regional Innovation Engines program (which is the largest place-based R&D investment since the land-grant acts). In July 2026, the program funded the UMKC-led Critical Materials Crossroads Engine in Kansas City, up to $160M over ten years. Workshop report.
Kauffman's flagship research calls on entrepreneurship and the economy. I helped design and run them: 1,100+ submissions, $12M deployed across 60+ projects. The 2018 grantee announcement, quote included.
Wrote the 2016 call for proposals that launched Kauffman's future-of-work research funding: before it was fashionable, long before it was urgent. The program's convenings and commissioned work ran for years and some related work continues to come out.
Contributor to Kauffman's bipartisan entrepreneurship policy platform, built with the Start Us Up coalition of 130+ organizations: a roadmap for reducing barriers to starting a business. The plan.
Rooms I've worked: OECD, Brookings, NSF, NBER, the U.S. Senate, the Federal Reserve Bank of Kansas City, and the Urban Institute, among others. Ninety-plus events as speaker, chair, or moderator, and the job in every one of them was the same: translate between the people who produce knowledge and the people who need it.
Talks, forums, and interviews. In 2026 I interviewed U.S. Senator John Hickenlooper for the national relaunch of the Kauffman Indicators, and built the briefing and interview protocol for former SBA Administrator Isabel Guzman.
Writing about AI and work is one thing. Building things with AI and seeing what holds up is a different thing. Everything below is just a prototype and should be treated as such. This website is one of the exhibits.
Who is actually adopting AI: businesses, workers, states. Built on published Census, Fed, and Anthropic data.
46 volunteers, 60 years of Kauffman history, and an AI tool built for real needs.
A toy model based on the points in my Liquidity Crisis of Expertise blog post.
A running list of work now possible thanks to frontier LLM models.
Apple pie, calzones, simit, and breads. Human-AI collaboration (AI recipe, my human hands and patience).
The site you're reading is itself a Lab project. Much easier to build these days (looking at you Angelfire, Geocities, and all the other platforms I used in the past).
A math game built for my daughters to use this summer, but available to all. They thought it was "kinda neat".
A game where you get to run a covert agency through differnt time periods (Cold War to now). Strategy game influenced by some of my favorite games as a kid like Covert Action, SimCity, Conflict, and Age of Empires II.
I grew up in Nevada, Missouri (pronounced Knee-vay-duh because it's a rural Missouri town!).
I studied economics at Truman State, then spent five years in market research and managing healthcare certification accounts. Two years as Director of Research at the Kansas City Business Journal.
The rest of my career has been at the Ewing Marion Kauffman Foundation, in two stints (2006–2008 and 2015–present). I've worked at the intersection of entrepreneurship research, labor markets, inclusive prosperity, and the economics of AI. Along the way came an executive MBA from Washington University's Olin School, a seat on the Foundation's core AI working group, and the Legacy Project, which turned 60 years of institutional history into searchable resource.
One constant is translation. Researchers and decision makers rarely speak the same language. My role is to connect the two so that research can be actionable.
Agents and Agency is where my current writing lives. The Lab is where I experiment with what AI tools can do, so that I'm not just writing in the abstract.