AI for Australian NFPs — assistance, not autonomy
Practical AI built into the workflows your staff already use. Hosted in Australia on Azure OpenAI. Audit-trail by default. Designed for organisations that have to defend every automated decision to a funder, a board, or a regulator.
The principle: AI assistance, not AI autonomy
Every AI feature in CRM for NFP suggests — staff approve. We don't auto-send messages, auto-close cases, or auto-allocate funding. Three reasons:
- Funder accountability. If a DEX-reported outcome was generated by AI without a human in the loop, you can't defend the data when audited.
- Client safety. Vulnerable-client triage involves judgement an LLM can't make.
- Regulatory drift. AI regulation is moving. Staying on the "AI-assists-humans" side of any future line is the safer bet.
Every AI suggestion is logged with the prompt, the response, the staff member who actioned it, and the outcome — the same audit trail as every other action in the CRM.
Three AI use cases shipping today
AI-assisted enquiry triage
When a new enquiry lands — web form, email, phone notes — AI proposes the program area, urgency, and best-fit staff member. Your team approves or overrides in one click. Reduces routing time and gets vulnerable clients to the right intake worker faster.
Case-note summarisation
For long-running cases, AI generates a one-paragraph summary of recent activity, flags overdue actions, and surfaces previous similar cases. Useful when a client is reassigned or when a manager reviews a case for the first time in months.
Follow-up suggestion
After an interaction is logged, AI suggests the next step (booking, referral, follow-up call) based on past patterns. The suggestion appears as a task draft your staff member approves — never sent automatically.
Where the AI runs — and where it doesn't
- Azure OpenAI Service — Microsoft's enterprise AI service, not the public OpenAI API. Different commercial terms, different data-handling guarantees.
- Australian datacentres — deployed to Australia East (Sydney) by default. Your client data does not leave Australia.
- Not used for model training — Azure OpenAI's commercial terms exclude your data from training Microsoft's or OpenAI's models.
- Retention controls — prompt and completion logs are retained per your organisation's policy, not Microsoft's default.
Compliance posture
For NFPs under DEX, ACNC, NDIS, or CHSP reporting obligations, AI introduces a defensibility question: can you prove a human made the decision? Our answer:
- Every AI suggestion is recorded as a draft, not an action.
- The audit log captures the prompt, the suggestion, the staff member who approved or rejected it, and the timestamp.
- A funder or auditor can trace any reported outcome back to the human decision that produced it.
If your sector's regulatory line moves — particularly around automated decision-making in vulnerable-client services — the audit trail is already there.
See AI assistance in action
A 30-minute walkthrough of the three use cases above, using anonymised NFP scenarios. No pitch deck, no AI hype — just whether this fits your workflow.