Fast Forward‘s 2025 AI for Humanity Report offers a snapshot of how nonprofits are using AI, dives into the emerging sector of AI-powered nonprofits (APNs), and provides takeaways for nonprofits.

The report focuses on the following categories of nonprofit AI use:

• AI-Powered Nonprofits: Organisations that develop AI solutions as a core part of their social impact work. Examples include AI-powered chatbots, recommendation engines, or systems that organise resources for end users. Importantly, end users may not always interact with the AI directly, but it is central to the nonprofit’s impact.
• AI-Assisted Nonprofits: Organisations that leverage AI behind the scenes to improve efficiency. Common use cases include marketing, grant writing, and automated workflows. While these applications do not directly change the product or service an end user receives, they free up capacity and indirectly enhance outcomes.
• Both: A subset of nonprofits reported using AI in both ways and are included in both categories above.

Section 1: AI-Powered Nonprofits

AI-powered nonprofits (APNs) represent one of the most promising fronts in social impact. APNs build AI into the core of their programmes, developing tools like chatbots, recommendation engines, and translation services to directly serve beneficiaries. In other words, AI isn’t just an efficiency layer for these organisations; it is the mechanism through which impact is achieved.

Because the field is still emerging, much of this work is early-stage. In fact, 40% of APNs surveyed have been using AI for a year or less. Behind every launch are unseen but critical phases, like sourcing and cleaning data, developing benchmarks, fine-tuning models, and running countless experiments. The costs of these early efforts are steep: nearly half (48%) of APNs cited cost of implementation as their biggest barrier. Yet with many funders waiting to see measurable outcomes before investing, APNs often face a catch-22: needing capital to prove impact, but needing proven impact to unlock capital.

Despite these headwinds, the results are already remarkable. One respondent reported that AI enabled them to quadruple their annual capacity for translations. Another saw response times for end users decrease by more than 75%. These stories illustrate how, when funders support experimentation, the payoff for communities can be transformative.

Top Issue Areas of AI-Powered Nonprofit Respondents:

• Education
• Economic Empowerment
• Health

AI Tenure: How Long APN Respondents Have Been Using AI at Core of Solution:

• <6 Months: 12%
• 6 Months-1 Year: 28%
• 1-2 Years: 28%
• 2-5 Years: 20%
• 5 Years: 13%

Key Finding 1: Nonprofits are building AI solutions at every size and stage

AI-powered nonprofits are showing that innovation isn’t confined to well-resourced Silicon Valley startups or large R&D labs. Half of surveyed APNs have 10 or fewer employees, well below the U.S. national average of 42 employees per nonprofit. Nearly a third (30%) of APNs have budgets of £500K or less. These lean teams are nonetheless building and deploying AI solutions that reach thousands — even millions — of people.

Employee Size:

• 1-10 People: 48%
• 11-50 People: 36%
• 51-200 People: 10%
• 201+ People: 6%

Key Finding 2: More funding is critical to unlock AI’s potential

Nearly half (48%) of APNs report that adopting AI has raised their expenses, and 84% say additional funding is most needed to continue developing and scaling their work. Building effective AI requires upfront investment in staff, infrastructure, and validation, long before the benefits show up for beneficiaries.

But that investment can pay off. At the smallest budgets, APNs are serving thousands, a median of just under 2,000 lives. At £1M budgets, median reach jumps to half a million people. And at £5M+, APNs are reaching millions, with a median impact of 7M lives. With resources, pilots can become sector-shaping tools.

Median Lives Impacted by Budget:

• <£100K: 1,852
• £100-500K: 40K
• £500K-1M: 125K
• £1-5M: 500K
• £5M+: 7M

Key Finding 3: APNs are embedding community voices into AI design

With government regulation lagging, AI-powered nonprofits are stepping up to model what responsible AI can look like. Their practices go beyond compliance: 61% of surveyed nonprofits customise large language models (LLMs) with their own data — a step that can help them tailor tools for the specific communities they serve. Plus, 70% regularly incorporate community feedback into system updates. By centring community voices in development, APNs ensure their tech reflects lived realities and builds trust with the people they serve.

The Good News? 71% of APNs already have processes in place to assess and mitigate risks. Still, the message is clear: nonprofits know the potential AI holds, but they are equally focused on ensuring it is deployed safely, equitably, and without exacerbating existing inequities.

Key Finding 4: APNs need technical expertise to match their ethical ambitions

Even with these proactive measures, concerns remain about safety and responsible use — a tension that shapes the way APNs approach adoption. Data privacy is a top challenge in adopting AI, cited by 48% of respondents. Another 41% pointed to the lack of in-house technical expertise, underscoring that without dedicated experts, it can be daunting for nonprofits to confidently assess whether their systems are truly ethical or secure. To close this gap, nonprofits must prioritise technical expertise as part of their AI strategy, and philanthropy should step up to make those hires and capacity-building efforts possible.

Key Finding 5: Most APNs start with chatbots, but time and experimentation will unlock bigger opportunities

Chatbots are the entry point for many AI-powered nonprofits — 67% use them today. But experimentation is quickly expanding into areas like content personalisation (51%) and research assistance (32%). Tenure also shapes the depth of adoption: 65% of APNs with under two years of experience rely on ready-made or lightly customised tools, while 63% of more seasoned APNs have moved on to developing fully in-house solutions. Today’s early experiments are only the beginning.

Current AI Applications:

• Chatbots: 67%
• Content Personalisation: 51%
• Research Assistant: 32%
• Translation Services: 29%
• Monitoring & Analytics: 28%
• Image/Voice Recognition: 26%
• Other: 28%

Key Finding 6: Collaboration among nonprofits could multiply AI’s impact

Collaboration is emerging as a powerful force among AI-powered nonprofits. Today, 43% of APNs have already made their tools fully or partially open, signalling a strong culture of sharing. Maturity also matters: organisations with more than two years of AI tenure are significantly more likely to open source or share their tools (53%) compared to newer APNs (38%).

This willingness to share tools, data, and practices reduces duplication and accelerates collective learning across the sector. By building on each other’s progress, nonprofits can avoid reinventing the wheel and move faster toward scalable, community-driven solutions.

Emerging Risk:

AI-powered nonprofits rely heavily on public data to train models and scale tools affordably. 77% of APNs rely on public datasets, making them the most-used source overall. But in the U.S. specifically, many of the public datasets nonprofits depend on are at risk of disappearing.

Without sustained access to high-quality, open data, the progress of the APN sector could stall, undermining the tools communities rely on. Public data must be treated as essential digital infrastructure: protected, maintained, and expanded to ensure that nonprofits can continue to build equitable, impactful AI solutions.

Section 2: AI-Assisted Nonprofits

AI isn’t just powering nonprofit programmes — it’s also reshaping how nonprofits run behind the scenes. The majority of organisations in our survey (82%) are using AI for internal operations, applying it to everything from grant writing to content creation to workflow automation.

For lean teams, AI tools can be game-changing. Smaller nonprofits report the highest rates of staff-wide adoption of AI, a reflection of how much efficiency matters when every person wears multiple hats. Across the board, AI is helping nonprofits reclaim hours and free up staff to focus on mission delivery.

Still, adoption remains early-stage. Most AI-assisted nonprofits are starting with off-the-shelf LLM tools, and few have formal policies to guide responsible use. As internal AI use grows deeper, the sector will need more than efficiency. Organisations will need investment, training, and shared standards to ensure AI is used responsibly and equitably.

Top AI Use Cases:

• Grant writing & applications: 77%
• Content creation & marketing: 77%
• Data analysis & reporting: 59%
• Donor engagement & communications: 58%
• Workflow automation: 49%
• Internal knowledge management: 45%
• Volunteer coordination & communications: 23%

Percentage of Nonprofit Staff Regularly Using AI in Their Work:

• 81-100% of staff use AI: 48%
• 61-80% of staff use AI: 17%
• 41-60% of staff use AI: 10%
• 21-40% of staff use AI: 16%
• 0-20% of staff use AI: 9%

Key Finding 1: AI-assisted nonprofits are using AI to reclaim time for what matters most

Nonprofits are putting AI to work on the tasks that once consumed staff time: content creation and marketing (77%), grant writing (77%), and data analysis (59%) top the list, followed by donor engagement (58%), workflow automation (49%), and knowledge management (45%).

What makes this significant isn’t efficiency for its own sake, but what it makes possible. Teams report drafting entire grant proposals in a fraction of the time, using AI to turn bullet points into first drafts, or summarising complex documents in minutes instead of hours. One leader described saving “two or three hours just with document translation,” while another shared that building a web app that used to take three months now takes just an afternoon. These shifts free up staff to focus on what matters most: serving their communities.

Key Finding 2: Smaller nonprofits are leading on back-office AI adoption

The smallest organisations are moving fastest when it comes to AI for internal operations. Nonprofits with budgets under £100K report the highest rates of staff-wide adoption, with an average of 82% of employees using AI regularly. Among mid-sized groups (£100K–£5M), about 67% of employees use AI, compared to 59% at large organisations (£5M+).

Their flexibility and urgency to save time mean small nonprofits are often quicker to bring AI into daily work. With every staff member wearing multiple hats, AI becomes a daily copilot that helps with tasks and frees up precious time for mission delivery.

Key Finding 3: Internal AI use starts simple, but is poised to deepen

For now, nonprofits are leaning on off-the-shelf AI tools (89%) to save time on everyday tasks. Few are customising or building in-house, and that’s appropriate for back-office work. But with training and resources, today’s convenience tools could evolve into deeper integrations that help reshape how nonprofits run.

Which best describes how the AI tools your organisation uses for operations are built or sourced?

• Off-the-Shelf LLM AI tools (e.g., ChatGPT, Google Gemini, Anthropic’s Claude): 89%
• AI add-ons to existing tools (e.g., Canva Magic Studio, Notion AI, HubSpot Breeze): 57%
• Customised versions of LLM tools using our own data (e.g., Gemini Gems, custom GPTs, Claude Projects): 40%
• Tools developed completely in-house: 19%

AI Tenure: How Long Respondents Have Been Using AI for Operational Support:

• 6 Months – 1 Year: 34%
• 1–2 Years: 39%
• 2–5 Years: 13%
• <6 Months: 11%
• 5 Years: 3%

Key Finding 4: Responsible AI practices lag among AI-assisted nonprofits as opposed to AI-powered nonprofits

When it comes to responsible AI, nonprofits using AI only for internal operations are well behind their AI-powered peers. Just 35% have a public AI policy or one in progress (vs. 84% of nonprofits that are AI-powered only), 39% have risk mitigation processes (vs. 75%), and 44% use privacy controls like anonymisation (vs. 69%).

Part of the gap is perception: back-office uses like content creation and marketing may feel “low risk.” But as tools advance and nonprofits apply AI more broadly, the risks — from data privacy to bias — grow too. Establishing simple templates for policies, privacy practices, and risk mitigation could help raise the baseline of responsible AI across the sector.

Section 3: Takeaways for Nonprofits

AI can feel overwhelming, but nonprofits are already proving it doesn’t take massive budgets or deep tech teams to get started. The key is to begin with what matters most to your mission, and build responsibly from there. These takeaways highlight the moves that set nonprofits up for lasting impact. Start with AI for operations, but start responsibly.

For more actionable tips, how-tos, and case studies, visit our Playbook on AI for Humanity.

Start with AI for operations, but start responsibly

Many nonprofits with small budgets and teams are already improving operations with AI. Begin with ready-made tools for tasks like writing, reporting, or donor engagement. But even at this early stage, set simple guardrails, like privacy practices, transparency with staff, and basic policies, so that as your AI use grows, you’re building on a responsible foundation.

Plan for growth and its costs

Building AI into the core of your solution can help you reach more people, but scaling requires investment in staff, systems, and infrastructure. Factor in rising expenses early, and communicate clearly with funders about the support you’ll need to expand responsibly.

Put your community at the centre

No off-the-shelf model will know your beneficiaries as well as you do. The most impactful nonprofits customise AI with their own data and regularly incorporate community feedback. Building with your community — not just for them — ensures your tools are trusted, equitable, and reflect lived realities.

Build skills as you go

As your use deepens, prioritise investing in technical expertise, whether through new hires, fellows, or upskilling your existing team, so you can confidently assess risks and shape AI to your mission. Consider also building tech capacity into your board or advisory structures, ensuring you have trusted guidance as the technology and your adoption evolve.

Share, learn, and experiment

AI is moving fast, and no nonprofit should go it alone. Many APNs are already open-sourcing tools or sharing practices, creating a culture of collaboration. By experimenting, sharing lessons, and adapting ideas from others, you not only accelerate your own learning but also multiply impact across the sector.