John Rice / Minnesota AI Field Notes
Practical AI.
A builder’s
field notes.
Guides for Minnesota business owners choosing a first AI project or helping their team learn. Alongside them: OpenClaw, agent systems, and the decisions behind the software. Practical examples, clear sources, and resources you can use.
Compare the whole cost.
What changes the cost and scope of an AI consulting project?
Compare deliverables, dependencies, recurring costs, and acceptance conditions before approving an AI engagement. Understand what a proposal includes and what remains your responsibility.
Know what you are buying.
How to choose an AI consultant for your business.
Ask for evidence that matches the work, clear access and review boundaries, and a handoff your team can use. A checklist for comparing any provider.
Name the missing piece.
Use your existing AI tools or build a custom workflow?
Test the complete task before buying more software. Compare existing tools, process changes, and a contained build using review effort and operating requirements.
A useful first project.
Choosing a first AI project for your Minnesota business.
Compare three real tasks, check the inputs and review burden, and choose a contained project your team can evaluate.
Leave with a plan.
Plan an AI workshop your Minnesota team can put to work.
Build a working session around one approved task, hands-on review, and an operating guide staff can use after the session.
Put AI to work.
How to get your business up to speed with AI.
A practical starting plan for owners and operators: understand current use, choose one task, set data boundaries, train the team, and measure the complete job.
Clearer work.
Three OpenClaw workflows I would test in a business.
A client report, a quote follow-up queue, and a founder brief: three proposed workflows with clear inputs, review steps, and ways to judge the result.
A clear decision.
An AI pilot needs a scorecard before a demo.
Define the decision, record the full task, and separate an impressive draft from a workflow your team would choose to use.
Reviewable work.
Build an agent team around work you can review.
How I think about requirements, specialist roles, coordination, and human review across the AI Agent Playbook, ClawPort, and TicketToPR.
Check the evidence.
AI search visibility starts with evidence.
Choose customer questions, inspect mentions and cited sources, then make changes you can explain. A practical approach informed by building Peekaboo.
made usable.
Turning a specialist’s method into software.
Inside Your Post Coach: how a focused coaching method becomes feedback a writer can use, with Tony Albrecht’s expertise at the center.
with evidence.
What a meeting assistant needs to show its work.
Inside an agency assistant prototype: source references, unresolved questions, and the review steps behind a meeting brief.
understood.
Before you automate it, understand the work.
A practical way to choose your first AI workflow without turning a small experiment into a systems overhaul.
handoff.
Your agent needs a handoff, not a pep talk.
Give an agent a clear destination, enough context, and boundaries it can actually follow.
repeatable.
Build a prompt you can use again next week.
Turn a promising one-off prompt into a small working tool with examples, checks, and an owner.