In an increasingly digital world, non-profit organisations’ staff and volunteers face mounting pressure to optimise their fundraising efforts. Limited budgets and competition for donor attention and funding (particularly following USAID cuts) are making it more challenging than ever to sustain vital projects. Fortunately, advancements in artificial intelligence (AI), particularly large language models (LLMs), offer transformative opportunities for NPO teams to enhance their fundraising strategies. These sophisticated tools, capable of generating human-like text, analysing data, and automating tasks, can assist fundraisers in connecting with supporters, streamlining operations, and maximising income.
Luckily, one doesn’t need to have advanced technical knowledge to utilise AI. One just needs to be aware of how to use it to assist in specific needs. It is vital to know the basics – particularly to assist in saving time. Large language models are superb for rapid online research. Most people dipped their toes into this new world via ChatGPT. Google is so 2020! Updates to LLMs are continually being released, and new programs are appearing apace. Grok is my current favourite. (I recently asked it where I could access certain documents. Not only did it instantly tell me what I needed to know, but, as I had to write an email with specific information, it also produced the wording!)
LLMs are excellent at summarising long documents. And one of their greatest properties is error-free English writing. (Don’t forget to request British English, unless American English is specifically required.) This is an immense help in many aspects of fundraising writing. However, humans must stay in control though and ensure that it’s not obvious that an LLM was used. Giveaways include numbered or bulleted paragraphs, a ‘conclusion’ or ‘in summary’ paragraph and, embarrassingly: simply copy/pasting the question at the end about whether the information was helpful!
One of the most immediate ways LLMs can support non-profits is by helping to create persuasive and emotionally resonant fundraising appeals. Writing effective donation requests requires a delicate balance of storytelling, clarity, and a strong call to action – skills that LLMs excel at. By providing a few key details about an NPO’s mission, target audience and project goals, non-profits’ fundraisers can use LLMs to generate tailored emails, social media posts, website content or crowdfunding copy. AI tools are also becoming better and better at generating images.
LLMs can also A/B test different versions of appeals to individuals (via email or social media) by generating multiple drafts with varied tones – urgent, hopeful, or informative – allowing one to identify what resonates most with donors. This data-driven approach, typically resource-intensive, is now rapidly accessible.
Personalisation is key to building lasting relationships with individual donors, yet many NPOs lack the capacity to customise such communication at scale. LLMs can analyse donor data – such as past donations, interests, or engagement history and generate personalised messages that feel authentic and relevant. For example, an environmental NPO might use an LLM to thank a long-term supporter: ‘Dear Sarah, your R500 donation last year helped us plant 30 trees in the Eastern Cape. Thank you for being a champion of our planet.’
This level of personalisation fosters a sense of connection, encouraging repeat donations. LLMs can also segment audiences based on demographics or giving patterns, producing tailored emails/updates that speak directly to different groups, such as high net worth donors, regular givers, or first-time supporters. By automating this process, fundraisers can maintain meaningful engagement.
Securing corporate or local (or foreign) trust donations is a lifeline for many non-profits, but the application process is often time-consuming and requires polished, persuasive writing. LLMs can streamline this by drafting initial proposals based on an NPO’s mission, project details, and funding requirements. For example, a fundraiser seeking money for a high school afternoon project, could input basic information – area, number of learners, and budget – and receive a coherent first draft: “Our centre in Khayelitsha aims to support 150 at-risk teenagers with free after-school activities, requiring R200,000 to cover staff, meals and resources.”
While human oversight is essential to refine these drafts and ensure accuracy, LLMs reduce the initial workload, allowing staff to focus on strategy and relationship-building with funders.
Social media is a powerful and ever more vital individual fundraising tool, yet maintaining an active, engaging presence can be daunting for understaffed NPOs. LLMs can generate a steady stream of posts, from awareness-raising content to donation prompts, tailored to platforms such as X, Instagram, or Facebook. For example, a mental health organisation could request a week’s worth of posts, and an LLM might produce: “One in four people in South Africa face a mental health challenge. Your R30 donation can fund a vital helpline call.”
Beyond content creation, LLMs can analyse past post-performance – identifying which topics or formats (e.g., videos, polls) drive the most likes, shares, or donations – and suggest strategies to boost engagement. They can also respond to simple follower queries with pre-approved messages, freeing up staff to handle more complex interactions.
Understanding donor behaviour is crucial for refining fundraising strategies, yet many non-profits lack the tools or expertise for in-depth analysis. LLMs can process large datasets – such as donation records or survey responses – and summarise trends in plain language. For example, an LLM might report: “In 2024, donations peaked in June, with 60% of gifts over R250 coming from donors aged 50 and over.” These insights enable fundraisers to target campaigns more effectively. While LLMs can’t replace dedicated data analysts, they provide a cost-effective starting point for organisations with limited resources.
LLMs offer immense potential. But non-profits’ staff and volunteers must navigate ethical considerations. Over-reliance on automation risks de-personalising donor relationships. Human oversight remains essential. Data privacy is vital and NPO staff must ensure donor information is handled securely and complies with the POPI Act. Additionally, LLMs occasionally ‘hallucinate’ (produce inaccuracies or generic content), requiring staff to review outputs carefully.
South Africa’s go-to NPO lawyer, Ricardo Wyngaard will be presenting a vital talk on NPOs Using AI – the Legal Considerations at the Annual Online Papillon Fundraising Conference on 20 & 21 May.
