Skip to content
Make Sense Of It

Blog

Insights, thinking and updates on AI for social impact.

Why we updated our playbook for charities

A year ago we published the AI Playbook for Charities. It was already going out of date. We’ve just published the second edition.

Edd Baldry & Suzanne Begley

Why PDFs give AI a headache

“Just upload the PDF” is the most misleading advice in charity AI. The gap between what that implies and what it takes to reliably extract data from real documents is where most of the interesting build work lives.

Edd Baldry

What 84 AI recipes taught us about the DIY boundary

We’ve mapped 84 practical AI use cases for charities. The most useful thing about the collection is what it reveals about where self-service stops and custom development starts.

Edd Baldry

  1. Navigating GDPR and large language models: a practical guide for charities

    Can your charity send supporter data to an American API? The answer is yes, but the detail matters. Here’s how to get it right.

    Suzanne Begley

  2. Data quality is the real AI strategy

    After two years of working with charities on AI, the single biggest predictor of success isn’t AI capability – it’s data quality.

    Edd Baldry & Suzanne Begley

  3. The real cost of AI for charities – and why custom is cheaper than you think

    Cutting through the “AI will save you money” narrative. When AI saves money, when it costs more, and why the economics of build vs subscribe have shifted.

    Suzanne Begley

  4. Claude Code, agentic AI, and what METR’s data means for charities

    AI autonomy capabilities are doubling roughly every seven months. That pace matters even if your charity isn’t ready to build anything yet, because the economics of what’s possible are shifting underneath you.

    Edd Baldry

  5. The Charity Digital Skills Report 2025: 76% of charities now using AI

    AI adoption is up 15% in a year. Charities have gotten started, but the governance, board capability, and strategy to support it are still catching up.

    Suzanne Begley

  6. The limits of Copilot and Power Apps: when off-the-shelf AI isn’t enough

    Copilot and Power Apps are fine for what they do. But charities are discovering the ceiling, and the problems that matter most sit above it.

    Suzanne Begley

  7. AI isn’t a computer. It’s a bridge.

    Someone asked us on stage recently whether it was safe to use AI to translate content into other languages. They wanted a yes or a no. That’s the wrong kind of question for this kind of technology.

    Edd Baldry

  8. Why charity meeting notes are deceptively hard for AI

    “Just transcribe the meeting” sounds simple. For charities, it’s anything but.

    Suzanne Begley

  9. Why we wrote a playbook for charities to use AI

    Almost everybody we spoke to said we shouldn’t do it. A small agency didn’t have the bandwidth. We did it anyway, because nobody was bridging the gap between what San Francisco was saying and what the third sector actually needed.

    Suzanne Begley

  10. Thinking about AI maturity

    It’s hard to be a beginner. But admitting you’re a beginner is the only honest place to start, and having a framework for what progress looks like makes it easier to have the right conversations.

    Edd Baldry