Logo

For most charities and membership organisations, the challenge with AI is in knowing where to start in a way that delivers real value quickly, without creating unnecessary risk.

In our latest webinar, AI in Practice: Building the Foundations to Get Started Safely, one insight stood out: most organisations are already experimenting with AI, but very few are seeing meaningful impact. The differentiating factor is how the technology is being applied.

The organisations seeing real results aren’t trying to roll out AI everywhere. They start with specific problems, test low risk use cases and keep the human in the loop throughout.

The real issue: activity without direction

Across the sector, AI adoption is happening in pockets. Teams are experimenting, and people are figuring out tools as they go.

Without a clear approach, this leads to a familiar pattern: lots of activity, inconsistent results, and limited organisational impact. AI starts to feel underwhelming, not because it isn’t powerful, but because it isn’t being applied with a strategy in mind.

The organisations that succeed do one thing differently: they identify a concrete problem and focus on solving it efficiently, rather than deploying tools everywhere at once.

Start where the risk is low and the value is immediate

The most effective starting points with AI aren’t complex or transformational. They’re practical, everyday tasks where the benefit is immediate and the risk is low.

Drafting content and first-pass outputs

AI can produce first drafts of emails, reports, funding applications, or internal documents in seconds. Humans review and refine the output, keeping risk low and accuracy high. The time savings are immediate, which builds confidence and momentum for further applications.

Summarising meetings and capturing actions

AI can take the friction out of meetings by producing clear, consistent summaries and highlighting key decisions and next steps. The real value isn’t just saving time, it’s ensuring that important details aren’t missed and everyone is on the same page.

(Quick tip from the webinar: if meetings are recorded or transcribed, make sure all participants give consent. It’s not just polite, it can be a legal requirement.)

Research and first-pass analysis

AI excels at summarising large amounts of information, mapping out topics, or producing initial structures to react to. This speeds up the slowest part of many workflows while keeping humans in control of the final output. AI handles the first 60–70% of the work, leaving teams free to focus on judgement, refinement, and decision-making.

Why some teams get far better results than others

One of the biggest differences we see is in the quality of prompts.

Generic requests produce generic outputs, while specific, well-structured prompts produce something far more useful.

For example, asking AI to “write a donor email” will get a basic, bland result. Adding context and detail– around the audience, tone, examples of what good looks like, and what to include or avoid– transforms the output into ready-to-use, high-quality content.

Small shifts in how AI is instructed can have a huge impact. Our webinar showed examples of prompts that produce results teams can actually use. Watch on demand to see these in action.

You don’t need perfect conditions to start but you do need guardrails

You don’t need a fully developed AI strategy to get started. But you do need some basic rules:

  • Keep humans in the loop
  • Be clear on what data can (and can’t) be used
  • Align internally on appropriate AI use

Human oversight is especially important for anything involving facts, statistics, legal or compliance information, or content shared externally. Many free AI tools aren’t designed for sensitive organisational data and using them without clarity introduces real risk.

With these guardrails in place, organisations can start safely, without slowing down progress.

The takeaway

Organisations seeing real impact aren’t doing more, they’re doing it strategically. They start with a specific problem, test low risk use cases, keep humans involved, and build evidence before scaling further.

AI doesn’t need to start with transformation; it needs to start with something that works, and with a clear-eyed understanding of what it can and can’t do.

This session is part of our wider AI webinar series, where we focus on practical, real-world applications of AI across the sector. If you’re ready to move beyond experimentation and start seeing measurable results, register for our AI Masterclass on 22nd April, check out our next webinar in the series, or get in touch with our team today.