Draft from verified facts
Write grant and report text using only supplied facts and flag missing evidence.
For charities and nonprofits
AI for nonprofits: practical uses, limits and staff training
AI for nonprofits is most useful for drafting communications, grant applications, meeting summaries and reports, provided staff check every claim, keep donor and beneficiary data out of unapproved tools and leave decisions with people. Training is what turns those rules into habits.
Six skills worth putting to work.
Keep impact claims honest
Remove outcomes, numbers and quotes the organisation cannot support.
Protect donor and beneficiary data
Use placeholders and templates instead of pasting records into AI tools.
Summarise governance material
Condense minutes or guidance for trustees and check the summary against the source.
A grant summary for a fictional reading charity
The Riverside Reading Project is fictional. Facts: weekly reading sessions in two primary schools, 18 volunteers, a request to fund a third school, and a parent survey in which most respondents said their child read more at home. Use an AI tool to draft a 200-word grant summary, then remove anything not supported by the facts. Submit the summary and a note listing what you removed.
Supported claims only
Every statement can be traced to the supplied facts; no invented numbers or outcomes remain.
Clear request
The funder can see what the money is for in one reading.
Gaps flagged
Missing evidence is marked for a person to supply rather than filled in by the assistant.
01 / For charities and nonprofits
How can nonprofits use AI?
Most of the practical uses are writing and summarising. NCVO, the UK membership body for voluntary organisations, lists communications such as press releases, emails and case studies; fundraising tasks such as drafting tailored grant applications and analysing donor data; governance tasks such as summarising meetings and condensing regulatory guidance; planning; and evaluation work such as designing surveys and producing reports. NCVO also notes that small charities lag behind larger ones, mainly because of limited funding, time and skills.
ChatGPT for nonprofits, and the other general assistants, help most where a staff member already knows what good looks like and needs a faster first draft. They help least, and can do harm, where the output is a factual claim about impact or a decision about a person.
02 / For charities and nonprofits
How do you use AI for a grant application without overstating impact?
Give the assistant only verified facts and tell it not to add outcomes. Here is a worked example from a fictional charity, the Riverside Reading Project. The supplied facts are: weekly reading sessions in two primary schools, 18 volunteers, a request for funding to add a third school, and a parent survey in which most respondents said their child read more at home.
The prompt: "Draft a 200-word project summary for a grant application using only these facts. Do not add numbers, outcomes or quotes that are not in the facts. Where the funder would want evidence we have not supplied, add a bracketed note."
A weak draft, from a prompt asking for a compelling summary, says the project has "transformed the lives of hundreds of children" and "doubled reading levels". Neither is in the facts. A funder who asks for the evidence will not find it, and the charity's credibility takes the damage.
The reviewed draft describes the sessions, the volunteers and the survey result in the survey's own terms, and adds a bracketed note: "[Add number of children reached and survey response count]". The writing is calmer and every sentence can be defended.
03 / For charities and nonprofits
What data should nonprofit staff keep out of AI tools?
Keep beneficiary records, safeguarding information, health details and donor databases out of any tool the organisation has not approved for them, and remove names before asking for a draft. A thank-you letter template does not need the donor list; a case study does not need the client's real name until a person has written the final version with consent.
The Charity Commission for England and Wales says trustees remain responsible for decision making and that it is vital this is not delegated to AI or based on AI-generated content alone. It highlights data security risks, especially under GDPR, and suggests charities consider whether an internal AI policy would be beneficial. Rules elsewhere differ, and this page is general information, not legal advice.
04 / For charities and nonprofits
What does a safe AI-assisted donor letter look like?
A safe one is written from a template with placeholders, not from pasted donor records. The staff member asks: "Write a 120-word thank-you letter for a regular donor to a children's reading charity. Use [first name] and [gift amount] as placeholders. Mention our weekly school sessions. Do not mention outcomes we have not supplied." The mail merge, run in the organisation's own system, fills the placeholders. The donor data never leaves it.
05 / For charities and nonprofits
What does AI training for nonprofits need to include?
It needs the same core skills as any team, briefing, drafting, checking and data boundaries, practised on the documents a nonprofit actually produces. Tosheo.ai is being built to provide that through a five-minute diagnostic, short role-specific lessons with hands-on tasks graded against visible rubrics, and practice scenarios. Managers see skill progress and completion, never chat logs. A final practical assessment leads to a recognised certificate with a verifiable URL.
Roles can be chosen from a preset list, such as Marketing for a communications officer or Ops for a programme coordinator; custom roles are planned for the Growth plan. Tosheo.ai is pre-build. Everything described here is planned launch scope, not a record of customer training, assessment results or outcomes.
Grants and reports
Draft from verified facts and mark every gap where evidence is needed.
Communications
Turn a programme update into a newsletter item or social post without inventing results.
Governance
Summarise a meeting or a long piece of guidance, then check it against the original.
Data boundaries
Use templates and placeholders instead of pasted donor or beneficiary records.
06 / For charities and nonprofits
When is Tosheo.ai not the right choice for a nonprofit?
If your organisation has fewer than five staff who would use it, the planned five-seat minimum may not make sense; the free guides on this site and a one-page AI policy may be enough. No nonprofit discount has been announced, so check the pricing page rather than assuming one.
It is also not the right fit if you need training on a specific donor CRM's built-in AI features, lessons for volunteers in languages other than English, or LMS integrations; none of these are included at launch. And because the product is pre-build, it cannot help a team that needs training this month.
Before you decide
A few good questions.
Can AI make my charity more successful?
It can make some writing and summarising work faster, which frees staff time. It cannot make decisions for trustees or replace evidence of impact, and this page makes no promise of fundraising or service results.
Which AI tools should a nonprofit use?
Start with a general assistant the organisation approves and can manage, such as ChatGPT, Claude, Gemini or Copilot, and check the current terms for the plan you choose. Specialist fundraising tools are separate purchases with their own data questions.
Is it safe to put donor data into ChatGPT?
Not by default. Use placeholders and templates instead, and only use personal data in a tool and plan your organisation has approved and that your data protection obligations allow.
Do charities need an AI policy?
The Charity Commission for England and Wales suggests charities consider whether an internal AI policy would be beneficial. A short policy covering approved tools, data rules and review is a sensible start for most organisations.
Is there a nonprofit price?
No nonprofit discount has been announced. The planned per-seat plans and the five-seat minimum are on the pricing page and may change before launch.
Keep exploring.
AI policy for small organisations
Take a closer look ↗Writing an AI acceptable use policy
Take a closer look ↗Is it safe to use ChatGPT with company data?
Take a closer look ↗Drafting and editing with AI
Take a closer look ↗AI training for small teams
Take a closer look ↗Planned Tosheo.ai pricing
Take a closer look ↗A useful next step
Give your team a path, not another playlist.
Explore the planned role-specific lessons, practical tasks and manager view.
Start free trial7 days free · Card required · Up to 5 seats