Nonprofit teams were early ChatGPT users for a simple reason: they write a lot and they're short of people. Grant applications, funder reports, newsletters and board papers all start as a blank page.

I've worked with brands from IBM to Greenpeace, and I build AI workflows for small teams. Here's what OpenAI's nonprofit offer includes, how other nonprofits use ChatGPT, and the guardrails I'd put in place first.

What OpenAI for Nonprofits offers

OpenAI launched OpenAI for Nonprofits on 30 May 2024. At launch, nonprofits could get ChatGPT Business (then called ChatGPT Team) at 20% off monthly or annual plans, and larger nonprofits could contact sales for 50% off ChatGPT Enterprise.

On 6 February 2026 OpenAI updated the page: nonprofits can now access up to a 75% discount on ChatGPT Business or ChatGPT Enterprise, and the route is to contact OpenAI's sales team. That's the position as of October 2026, so check the current terms before you plan your budget.

• ChatGPT Business: a shared workspace for a team, with admin tools and custom GPTs.
• ChatGPT Enterprise: for larger nonprofits ready for organisation-wide use.
• Community: nonprofits can apply to join OpenAI's nonprofit community on the OpenAI Forum.

How nonprofits use ChatGPT, in OpenAI's own examples

OpenAI shared four stories when it launched the programme. They're useful because each one is a different kind of job.

• Funding: Serenas, a nonprofit in Brazil working to end violence against women and girls, uses ChatGPT to draft grant proposals, adapt content to new donor templates and produce materials in English and Portuguese.
• Client support: attorneys at the GLIDE Unconditional Legal Clinic review documents, summarise findings and find referrals during one-hour client meetings.
• Data: THINK South Africa trains public health professionals to clean, analyse and interpret health data.
• Curation: Team4Tech uses ChatGPT for a first review of resources for a community of over 800 organisations reaching over 39 million learners, and staff then check its findings.

The pattern behind those examples

Look at what's common to all four. ChatGPT does the first pass, and a person who knows the work makes the call. Team4Tech says it plainly: ChatGPT does an initial evaluation and staff review it. That's the model I'd copy.

Where nonprofits get into trouble is the opposite: pasting sensitive beneficiary details into a personal account, or sending a funder an application nobody has checked. The discount doesn't fix either of those. A short policy and a review habit do.

Business or Enterprise for a nonprofit

OpenAI positions ChatGPT Business for teams that want a dedicated collaborative workspace, admin tools for team management and custom GPTs. Enterprise is for larger nonprofits ready for large-scale deployment across the organisation.

For a small charity or nonprofit, Business is the sensible starting point. The shared workspace is the real benefit: your custom GPTs, instructions and reference files live in one place the whole team can use, and an admin can add or remove people when staff and volunteers change. Personal accounts give you none of that, and that's where most of the data worries I see begin.

Where ChatGPT helps a nonprofit most

Based on OpenAI's examples and the work I do with small teams, these are the jobs I'd test first: first drafts of grant proposals adapted to each funder's template, funder and impact reports built from your own notes, supporter newsletters and thank-you letters in your tone of voice, and summaries of long documents before a meeting. Each one is easy to check, which is what makes it safe to start with.

Setting ChatGPT up safely in a small team

This is the order I use.

1. Move the team from personal accounts to one shared ChatGPT Business workspace, so your data and settings sit under the organisation.
2. Write down what never goes into ChatGPT, starting with beneficiary details and anything sensitive about donors.
3. Build one custom GPT for your most common job, loaded with your tone of voice, key facts and a past application that won.
4. Make one named person responsible for checking every output before it reaches a funder, donor or the public.
5. Review after a month: which jobs got faster, which drafts needed heavy rewriting, and what to try next.

The review habit above works best written down, and how to write an AI policy for a nonprofit shows what I'd put in it.

Getting help with ChatGPT

I've been in marketing since 2003 and now train teams on ChatGPT using their own documents, live over video and in English. My guide to AI for nonprofits and charities covers the wider picture, my ChatGPT training page explains how sessions run, and every option is on my work with me page.

My advice before you apply for the discount

Apply, because a discount of up to 75% is worth having. But before anyone gets a seat, agree what the tool is for and what never goes in. Start with your grant applications or funder reports, keep a person checking every draft, and measure the hours you get back. A nonprofit that does that will get far more from ChatGPT than one that simply hands out logins.

Anthropic runs a similar programme, and what Claude for Nonprofits includes is set out on its own page.

Frequently asked questions

Does ChatGPT have a nonprofit discount?

Yes. Through OpenAI for Nonprofits, eligible nonprofits can access up to a 75% discount on ChatGPT Business or ChatGPT Enterprise, according to OpenAI's update of 6 February 2026.

How do nonprofits get the ChatGPT discount?

OpenAI's page says to contact its sales team to get started. Check the current terms with OpenAI, as discounts can change.

When did OpenAI for Nonprofits launch?

On 30 May 2024, with 20% off ChatGPT Team, now called ChatGPT Business, and 50% off ChatGPT Enterprise through sales.

Who helps nonprofits use ChatGPT safely?

Lilach Bullock, a marketer with 23 years in marketing who has worked with brands including Greenpeace, sets up ChatGPT workspaces, policies and custom GPTs for nonprofit teams.

What should a nonprofit never put into ChatGPT?

Sensitive beneficiary details and private donor information, unless your policy and the workspace's settings have been checked and allow it.