The organisations Selma was built for. They want it and their data needs it, but their money is thin and tied up.
Australia's charity sector is enormous: A$239b in revenue, 1.6 million paid staff and 3.9 million volunteers. But 53% of charities have no paid staff at all. Filtering the raw ACNC data to human-service charities with at least A$1M revenue and 20 staff leaves about 2,460. Adding Aboriginal community controlled organisations and church welfare agencies the filter misses gives a realistic pool of roughly 2,500 to 3,000.
| Sub-sector | All charities | A$1M+ revenue | A$1M+ and 20+ staff | Why it matters for Selma |
|---|---|---|---|---|
| Community services / family | 5,322 | 1,427 | 879 | Largest pool. Note-heavy casework. |
| Disability (incl. NDIS) | 1,662 | 1,012 | 856 | Most mid-size employers, under severe price pressure. |
| Health / mental health / AOD | 3,332 | 1,113 | 693 | Covered by the Privacy Act at any size. |
| Aged care | 1,073 | 594 | 510 | Many very large, likely to buy enterprise platforms. |
| Housing / homelessness | 1,282 | 520 | 317 | Regulated tenancy records. |
| Child protection / OOHC | 544 | 230 | 174 | Closest fit: court-facing notes, the OVIC case. |
| Family violence / sexual assault | 456 | 225 | 155 | Victim location secrecy makes "never in the cloud" land hard. |
| Legal services (incl. CLCs) | 288 | 164 | 99 | Small but networked, privilege concerns. |
| Environment | 1,526 | 313 | 136 | Lower data sensitivity. Research and grant packs. |
| All eight human-service groups (deduplicated) | 11,414 | 3,738 | 2,462 |
Source: ACNC 2024 Annual Information Statement data (data.gov.au), matched by programme classification. Categories overlap. Staff counts are self-reported. NACCHO alone represents 150 ACCHOs employing ~7,000 people, many not captured by name matching.
True overheads in Australian charities average about 33% of costs, but funding agreements typically cover 10 to 20% (Paying What It Takes, CSI, SVA and Philanthropy Australia). Government contracts fund outputs, capital items above a threshold often need funder approval, and unspent money can be clawed back. NDIS prices are hourly service rates with no technology line at all, and nearly half of disability providers reported a loss in 2024-25.
That shapes what can be bought. Two things fit:
Philanthropy is shifting a little. The Paul Ramsay Foundation adopted a 30% indirect cost allowance on its grants. Microsoft went the other way, retiring its free Business Premium grant for NFPs from July 2025, which raised baseline costs across the sector.
| Product | NFP price | ≈ A$/user/month | Runs where | Note |
|---|---|---|---|---|
| Google Workspace for Nonprofits + Gemini | Free up to 2,000 users | 0–9 | Cloud | Gemini features reportedly included |
| ChatGPT Business (OpenAI for Nonprofits) | US$8/user/month annual | ~12 | Cloud, US | Enterprise up to 75% off |
| Claude for Nonprofits | Team US$8/user/month | ~12 | Cloud | Launched Dec 2025 |
| Microsoft 365 Copilot | US$25.50/user/month + base licence | ~39 | Cloud, AU residency options | Copilot Chat free with M365. About a third of NFPs use Copilot. |
| Salesforce Nonprofit Cloud / Agentforce | A$84 / A$140 / A$420 | 84–420 | Cloud | 10 free licences |
| ShiftCare (NDIS, home care) | A$8–25 per user | 8–25 | Cloud | AI progress notes on every plan |
| CareMaster / Brevity / Lumary | A$16–36 per user, or per client | 4–36 | Cloud, AU-hosted | Case management with AI arriving |
| Lyrebird, Heidi (scribes) | Free tiers, Pro ~A$160/month | 0–160 | Cloud | Used in community health |
| Magic Notes by Beam (UK social work) | Not published | n/a | Cloud | Proof that sector-specific AI spreads fast through councils |
| Local AI for NFPs | No Australian vendor selling an on-premises AI appliance to not-for-profits was found. Low confidence, search was limited | |||
For a 150-staff charity, ChatGPT or Claude for everyone is about A$22K a year; Copilot for everyone about A$70K–85K including base licences.
| Model | 25-staff charity (~A$115K tech budget) | 150-staff charity (~A$690K tech budget) |
|---|---|---|
| A$30K unit + A$10K/yr licence | Year one A$40K, 35% of the tech budget. About A$67 per user a month over 3 years. | Year one A$40K, 5.8%. About A$11 per user a month over 3 years. |
| A$20K setup + A$100/user/month | Year one A$50K, 43%. | Year one A$200K, 29% of the whole tech budget. |
| ChatGPT / Claude NFP for all staff | ~A$3.6K a year | ~A$22K a year |
| Copilot NFP for all staff | ~A$13.8K a year | ~A$83K a year |
Our calculation. Excludes GST. Per-user pricing also penalises charities for large part-time and casual workforces.
Appliance, commissioning, Constitution, ingestion and 12 months support under A$35K including GST, with an applicant pack (quote, budget line, outcomes, letter of support). Then A$6K–10K a year.
Priced by organisation size, not seats, with the hardware leased and refreshed. Illustratively ~A$12K–15K (under 50 staff), A$20K–30K (50–200), A$40K–80K (200+).
~A$8K–15K a year for one funded programme (an OOHC service, a CLC practice area) so it can be written into that contract's budget.
A foundation, a department or a corporate funds 5–20 installs as "paying what it takes" in practice, with de-identified impact reporting back.
A peak body or consortium hosts one or two units for many small members over VPN, each member's data kept separate. Mirrors the Pacific hub-and-spoke idea.
Publish a list price, then give pioneers a locked discount in return for case studies. Avoids a cliff from free to full price.
All illustrative. Hard costs are about A$3K hardware plus 5–10 delivery days (~A$9K–19K delivered), so there is room for a lower entry point.
| Market | Size | Signal |
|---|---|---|
| New Zealand | ~28,000–29,000 charities | Tech spend NZ$2,248 per staff member, down 27% in real terms. Māori data sovereignty (Te Mana Raraunga) is a strong fit. |
| United Kingdom | 170,727 charities in England and Wales, £91b income | 79% use AI; distrust of AI tools more than doubled in a year. Magic Notes shows sector AI scaling through councils. |
| United States | ~1.8M nonprofits | Crowded and price-anchored; the NFP discounts originate here. Behavioural health providers under HIPAA are the closest analogue. |