AI in context.
Non-profit.
Grants, programmes, and supporter work, with more time for the mission.
The mission is the work. AI takes on the admin around it.
Opportunity map
Where we would start.
Find the right place to begin.
Illustrative positions, not measured results.
Compare value and feasibility before committing.
Measure: baseline effort and validated gainsWhere to start
Three starting points.
Grant applications and reports
Drafted from programme data; staff submit.
Supporter communications
Responses drafted; people approve.
Programme knowledge
Sourced answers for staff and volunteers.
Signature workflow
Grant report to submission.
Same work. A different way through.
Today
The current sequence.
- Collect data
- Compile
- Draft
- Review
- Approve
- Submit
With AI
AI where it helps. People where it matters.
- Collect dataAI assisted
- CompileAI assisted
- DraftAI assisted
- ReviewWorkflow step
- ApproveHuman approval
- SubmitWorkflow step
- AI assisted
- Human authority
People stay accountable for the outcome.
3 of 6 proposed steps use AI assistance.
Measure: hands-on minutes per task, including reviewEstimate time savings for this workflowUse your own volume and times
Include review, exceptions and corrections in the assisted time.
Weekly time difference = tasks × (minutes today − minutes with assistance) ÷ 60.
Sector requirements
What has to be true first.
Decisions about people stay with people
Eligibility and support keep a named owner.
Beneficiary data stays protected
Access governed, every action logged.
Evidence for funders
Reports traceable to their sources.