Nationflow AdvisoryFunding agreement scanner
Drop a contribution agreement in. Out comes a clean summary: reporting dates, eligible expenses, deliverables, and any clauses worth flagging to council. No more reading 40 pages to find a deadline.
AI Workflow Consulting
Practical AI for band offices. Not chatbots. Not hype.
Most “AI for government” pitches don’t survive a week inside a band office. The systems are too generic, the training is too long, and the value disappears the moment your one technical staff member moves on. The work I build is the opposite of that — small, focused tools that take repetitive load off real workflows.
What I mean by “practical AI”
AI is good at a narrow set of things: reading long documents, pulling structured information out of messy text, drafting first passes of routine writing, and connecting tools that don’t naturally talk to each other.
Used carefully, that’s enough to recover several hours per week in a small office — without changing how staff work and without forcing anyone to learn a new platform.
Used carelessly, it produces confident-sounding nonsense and adds risk. So every workflow I build has a clear human review step, a paper trail, and an off-switch.
Where AI does not belong (yet)
- Anywhere community data needs to stay strictly inside the office without third-party processing
- Final approvals on funding submissions, council reports, or legal documents
- Decisions about housing allocation, individual files, or anything affecting members directly
- Replacing staff who already know the community context better than any model will
Example workflows
Six things AI is genuinely useful for in a band office.
Each one is a real, narrow workflow — not a platform you have to adopt all at once.
Reporting drafts from raw notes
Project notes, photos, and invoices in. A first-pass quarterly or final report out. Staff edit instead of write from scratch — usually a 60–80% time saving on routine reports.
Renovation & maintenance intake
A simple form members can use from a phone. Requests get triaged, deduplicated, and added to a tracker. Nothing falls through the cracks because it never made it out of someone’s inbox.
RFP & grant scanning
New funding bulletins, RFPs, and program updates summarized weekly. You see what’s relevant to your community without anyone reading every newsletter.
Email & attachment triage
Long email threads with attachments turn into structured records: who said what, what was decided, what’s outstanding. Useful when handing files between staff or back to council.
Searchable document library
Existing folders — drives, scans, attachments — made searchable by what’s inside the documents, not just file names. Old reports stop being lost.
How a workflow project goes
Small, focused engagements. Not platform rollouts.
Workflow review
We pick one or two real workflows that are eating staff time. No assessments, no surveys. Just look at the actual work.
Prototype
I build a working prototype — usually within a couple weeks — so staff can see it in action before any commitment.
Pilot
Small pilot with one team. Outputs reviewed by humans every step. Fix what’s wrong; keep what works.
Handover
Documentation written for the next person. Off-switches, edit access, and clear instructions on what to do if it breaks.
Community data, handled carefully.
Every workflow has a clear story for where data lives, what gets sent to a third-party model, and what stays inside the office. Defaults lean private. If a workflow can’t be done without sending sensitive data to an outside service, I’ll tell you up front and we’ll decide together whether the trade-off is worth it.
If your community has data sovereignty requirements (OCAP, local policy, funder restrictions), those come first — and they shape the design before anything else.
- Clear documentation of where data flows for every workflow
- Local-first or private-deployment options when sensitivity warrants it
- OCAP-aware defaults: ownership, control, access, possession
- No training of public models on community data
- Off-switches and audit logs built in from day one
- Tools and credentials handed over to the community at end of engagement
Honest take: not every band office needs AI right now. Sometimes the right answer is a clean spreadsheet, a fixed naming convention, or one shared inbox. I’ll tell you when that’s the case instead of selling something fancier.
Curious whether one of these workflows would help your office?
Book a 30-minute review. I’ll be honest about what AI is and isn’t worth doing for you.