Over the weekend, I sat down with a cup of iced coffee, my trusty dog, and the AI Guide for Government from the U.S. General Services Administration’s Centers of Excellence. (Yes, I know how to have a good time.)
It’s a detailed blueprint for how federal agencies can responsibly adopt and govern AI, complete with flowcharts, frameworks, and acronyms that could double as minor planets. I’d never stumbled across this particular report before, so I timed how long it would take for my eyes to glaze over.
Surprisingly, they didn’t. What caught my attention wasn’t the bureaucracy. It was the resonance.
Much of what’s in that federal playbook mirrors what we’re already doing here in Van Buren and St. Joseph Counties, only with fewer committees, more coffee, and a tighter budget.
We are, curiously, aligned. It’s proof that our local AI strategy isn’t playing catch-up with Washington. It’s walking in step, just wearing slightly muddier boots.
Below are ten principles from the federal guide and how we’re translating them into real work. Some are in full swing, others are still sprouting roots. Together, they form the backbone of DICE, the Digital Innovation Collaborative Exchange: our shared digital utility across both counties.
Think of DICE as a joint power grid for AI, automation, analytics, communications, and data governance.
1. AI Belongs in the Work, Not in a Silo
Federal principle: AI should live inside real workflows, not in a detached tech office.
Our reality: We’re embedding AI directly into daily work: zoning, FOIA, HR, communications, and beyond. But unlike a traditional “AI department,” we’ve created a shared AI team through DICE that works alongside departments rather than above them.
Departments own their use cases and workflows. DICE provides the technical muscle, strategic coherence, and implementation support. It’s a hybrid model: centralized capacity with decentralized execution. The bureaucratic equivalent of dancing in formation.
2. Trustworthy AI Starts with Trustworthy Data
Federal principle: Reliable AI requires reliable, structured, explainable data.
Our reality: We’re rebuilding both county websites specifically to be AI-friendly, structured, machine-readable, and context-aware. We’re assembling a coherent, discoverable, authoritative data repository that allows AI systems to work with the same clarity and confidence as a human clerk with decades of institutional knowledge.
Information isn’t treated like static content anymore. It’s civic infrastructure. Without clean, governed data, AI is basically an overconfident intern making things up.
3. People First, Technology Second
Federal principle: Invest in your workforce, not just in software licenses.
Our reality: Every staff member must complete mandatory AI policy training before gaining access to any AI tools. This sets clear expectations around ethics, security, and appropriate use. After that, we provide hands-on workshops and practical how-to guidance to build real fluency.
Our goal isn’t to make everyone an AI expert, but to make everyone AI capable. In local government, that’s not a luxury. It’s fast becoming a civic neccesity.
4. Clear Structure Beats Bureaucracy (Every Time)
Federal principle: Use integrated teams, clear governance, and a central resource.
Our reality: We’ve built a structure with three interlocking components:
- AI Task Force: The doers and testers. They experiment, pilot, and bring real-world experience back into the system.
- AI Steering Committee: The governors and guardrails. They provide oversight, policy alignment, and accountability.
- DID/DICE: The enablers. We provide the infrastructure, technical guidance, and long-term strategy.
This three-part harmony keeps innovation close to those who use it while ensuring oversight doesn’t turn into overreach.
5. Data Before Tools: The County Knowledge Base
Federal principle: Start with a strong data foundation.
Our reality: Our County Knowledge Base is the quiet powerhouse of everything we’re building. It’s a structured, machine-readable library of authoritative county data: GIS, property tax records, permits, ordinances, minutes, financial data, infrastructure, and more.
Instead of burying critical information in PDFs and file shares, we’re creating a living, queryable knowledge layer that AI systems can actually trust.
It’s not glamorous work, but neither is pouring a foundation. Until the walls fall over.
6. Start Small, Prove It Works
Federal principle: Pilot, test, and build from early wins.
Our reality: We started with AI tools that summarize meetings, assist with writing and communications, and automate simple workflows. Then came chatbots and specialized assistants.
These early wins are small, visible, and intentionally practical. They build credibility across departments and create momentum without overwhelming anyone. Minimum viable magic.
7. AI Won’t Replace People — But It Will Redefine Their Work
Federal principle: AI replaces tasks, not jobs.
Our reality: We tell staff plainly: AI isn’t here to take your job. But the person using AI might.
AI is a tool that isn’t going away, and we view AI fluency as a civic responsibility. The point isn’t job loss; it’s transformation. We want employees focused on judgment, creativity, and impact, not on repetitive, low-value tasks. Let the AI worry about those.
8. Mission Attracts Talent
Federal principle: Government can’t always outpay industry, but it can out-meaning it.
Our reality: Being among the first rural governments to lead on AI gives us something money can’t buy: purpose.
Through DICE, we pool resources to attract and retain skilled people who might never have considered working in county government. And through our nonprofit arm in development, we’re creating the ability to offer more competitive compensation while preserving public service values.
We’re also working closely with Western Michigan University faculty and students through the Center for Excellence in Research and Public Service.
This isn’t just a technology strategy. It’s a talent strategy.
9. Responsible AI Means Clear Ethics
Federal principle: Ethics, accountability, and transparency must be designed in, not bolted on.
Our reality: AI ethics is central to our approach. Our policies clearly define how, when, and when not to use AI. We train users to understand its limitations as well as its strengths, and we revisit these guardrails regularly.
This isn’t about paranoia. It’s about trust. Public trust is earned, not assumed, and we intend to earn it.
10. Governance That Fits the Scale
Federal principle: Strong but flexible governance keeps innovation moving.
Our reality: Our AI Steering Committee ensures alignment, accountability, and transparency without drowning innovation in red tape. DICE provides the technical backbone, but departments remain the locus of power and execution.
We’re building governance that fits our scale, not a federal bureaucracy with rural resources.
The Bigger Picture
When you compare the federal AI framework and what’s happening here, you don’t see two different worlds, just two different zoom levels.
Where Washington drafts strategy, we’re field-testing it in real communities. Where others are still writing about AI readiness, we’re already living it.
We’re not proving that rural governments can catch up. We’re showing that they can lead; responsibly, practically, and yes, sometimes with a little bit of duct tape and determination.
AI isn’t the future of government. It’s the present. And in Van Buren and St. Joseph Counties, it’s already clocked in for work.