🗞️ AI for Non Profits Network: Weekly Briefing 01/20

The weekly digest from a network of non-profits interested in AI. What's in this week's Briefing: ChatGPTs move into ads; What Claude Code & Cowork mean for your non profit; and loads of resources.

In The Briefing this week:

1. 👀 What’s Caught our Eye: ChatGPT Joins the Advertising Revolution

2. 💭 Thought for the Week: We Tried Claude Code and Cowork - Here’s How It Could Support You.

3. 🌟 Interesting News

4. 🌐 From Across the Network

Last Chance to join our first Free AI for Non-Profits Network workshop on 28th January at 2pm (EST). We’ll discuss 2026 goals and how to implement safe and structured pilots in key areas of your organization across an informal and open peer learning process.

We’ll also hear from CEO of Net Impact, Karen Johns on their recent AI pilot and it’s impact and future plans. Other peers include non profits for older adults, journalism, microfinance, minority communities and faith based organizations.

Reply to this email or send us a message if you are interested in joining.

1) 👀 What Caught Our Eye: ChatGPT Joins the Advertising Revolution

The AI advertising landscape just shifted dramatically – what does this mean for non profits?

On 16 January, OpenAI began showing advertisements to ChatGPT’s 800 million weekly users. Free and Go tier users in the US now see sponsored content woven into their AI conversations. While there’s no self-serve platform yet, this marks a pivotal moment in how people discover organizations online.

This amplifies what we explored in an earlier edition on Answer Engine Optimization: the way people find your organization is fundamentally changing. Traditional search engines are being replaced by AI assistants that synthesize answers rather than list links.

The numbers tell the story. Google’s AI Overviews already reach 1.5 billion users with embedded ads. Microsoft’s Copilot advertisements achieve 73% higher click-through rates than traditional search ads.

This is a land-grab moment. There are no established playbooks, no experts with a decade of experience. Cost-per-thousand impressions currently hover around $3 - absurdly low compared to traditional digital advertising. Those prices won’t last.

What this means for your nonprofit

Revisit your AEO strategy. If you haven’t optimized how AI systems understand and present your mission, start now. When someone asks ChatGPT “where should I donate to fight food insecurity?” your organization should be part of that answer.

Monitor which platforms accept advertisers – the landscape is evolving weekly. Early movers in previous platform shifts built sustainable advantages.

Finally, experiment while barriers are low. Test small campaigns on emerging AI ad networks. The nonprofits that understand AI-native advertising today will be the ones donors find tomorrow.

We’ll delve deeper into what this means with full guides in future issues. If you’re already experimenting, drop us a message so we can spotlight you to the network.

2) 💭 Thought of the Week: We Tried Claude Code and Cowork - Here’s How It Could Support You.

In recent weeks, our LinkedIn feeds and general industry circles have been abuzz with Claude Code and Cowork. Dubbed as ‘an AI agent that actually works’, Claude Code and its newly released companion Cowork, are designed to perform complex, multi-step tasks. We decided to give it a go and see what the implications might be for the nonprofit sector.

Claude Code launched late last year as a command-line tool allowing developers to delegate entire coding tasks directly from their terminal. But Anthropic noticed something unexpected: developers weren’t just using it for coding. They were using it for everything else too - organizing files, processing data, creating documents, managing workflows.

This observation led to Cowork, launched last week as a research preview for Claude Max subscribers. Think of it as Claude Code for non-techies. You give Claude access to a folder on your computer, and it can read, edit, and create files within that space. Need to organize a chaotic downloads folder? Want to create a spreadsheet from expense screenshots? Have scattered notes or research that need to become a coherent report? Claude handles it.

What makes this genuinely interesting is the level of agency involved. This isn’t the back-and-forth of a typical chat interface. You give access to your files and folders, set a task, Claude makes a plan, executes it, and keeps you informed. You can queue multiple tasks and let Claude work through them in parallel. It feels less like using a chatbot and more like leaving instructions for a capable colleague.

How Developers Are Using It

Speaking with Niall, Head of Design at non-profit developers Whitelabel, reveals how fundamentally this changes the development workflow. “I use Claude Code as a primary pair programmer,” he explains, “not just for one-off code generation here and there as with traditional chatbots.” He’s adopted what he calls Plan Mode - using Claude to reason through architecture and trade-offs before writing a single line of code. “This reduces false starts and limits rework once coding begins.”

More intriguingly, he runs multiple instances of Claude Code in parallel across different Git worktrees, enabling simultaneous development of features without context switching. “I often run the same task multiple times in parallel”. “It’s more expensive, but it allows me to compare alternative implementations side by side, generate different UI options, and evaluate code quality.” Because AI models are non-deterministic, this parallel approach surfaces better solutions than any single attempt would yield.

Of course, there’s a framework in place when dealing with sensitive topics in the non profit space: “We treat Claude Code as a collaborator rather than an authority. Outputs are reviewed, adapted, and integrated into existing code standards.” He uses it across the full development lifecycle - planning, scaffolding, refactoring, reviewing pull requests, generating documentation. It’s not replacing human judgment; it’s amplifying it.

Uses Beyond Coding

I recently pointed Cowork at a folder containing two years of case study drafts, interview transcripts, and research notes scattered across inconsistent file names. Claude reorganized the entire archive by theme and date, created a master index, and flagged duplicates I didn’t know existed. What would have taken me days took twenty minutes.

For non-profit organizations perpetually stretched thin, the implications are substantial. A program manager with participant feedback scattered across surveys, emails, and handwritten notes could delegate the entire synthesis process. Grant reporting data spread across multiple systems could be assembled into structured drafts automatically. A development director could have Claude research foundation priorities across websites and draft tailored inquiry letters while they focus on relationship-building.

But let’s be clear about risks, because non-profits can’t afford to be cavalier with sensitive data. Cowork is a research preview, currently macOS-only for Claude Max subscribers. More significantly, Claude can execute potentially destructive operations like deleting files if instructed, and AI can occasionally misinterpret instructions. There are also concerns about “prompt injections” - attempts to manipulate Claude through malicious online content.

The appropriate response isn’t to shy away, but to approach thoughtfully. Start with low-stakes tasks in sandboxed folders. Don’t grant access to sensitive information until you’re confident in system behavior. Treat it like onboarding a new staff member- provide access gradually, based on demonstrated competence.

What’s most compelling isn’t the current capabilities but the trajectory they represent. As these systems improve, the gap between “I wish we could...” and “We’ve built...” will continue to narrow for organizations without technical departments. The organizations that will benefit most are those building comfort with AI tools now, even in their imperfect early forms.

This also means rethinking how you allocate human talent. If routine processing and compilation can be delegated to AI agents, your staff’s comparative advantage shifts entirely to relationship-building, strategic thinking, and the contextual judgment that requires deep community understanding.

For non-profit leaders

When asked about implications for non-profit teams, Niall is bullish: “From zero to MVP with Claude Code is so quick. It gets stuff done in one shot, doesn’t require deep prompting or multiple turns for small proof of concepts.” One moderately technical person could spin out multiple internal tools rapidly, testing ideas that would previously have languished in the “someday maybe” pile. That volunteer management system you’ve needed for years? The donor follow-up tracker your development team keeps requesting? Suddenly these move from six-month projects to weekend experiments.

So, some final thoughts for starting with Claude Cowork (when it’s more widely released): Start small and specific. Identify one recurring administrative task that consumes staff time without requiring sensitive data. Set up a test folder. Try delegating that task to Cowork. Document what works and what doesn’t. The skills your team needs to develop aren’t coding - they’re the ability to break down work into clear tasks, articulate requirements precisely, and oversee automated processes effectively. Those skills will serve you regardless of how quickly these tools evolve.

Remember, we still have a few places left for our first Free non-profit AI workshop on 28th January at 2pm (EST). We’ll discuss all of the above in a peer learning informal environment.

Reply to this email or send us a message if you are interested in joining.

3) 🌟 Interesting News

More Power Together Launches as AI Referral Network for Nonprofits (Non Profit Pro)

Major charities leave X after AI images controversy (Civil Society)

Tencent seeks AI developers to protect vulnerable users (Tech in Asia)

4) 🌐 From Across the Network

• Shutterstock GenAI Pro: From Core Features to What’s New - How recent enhancements to Shutterstock AI make it easier to create high-quality, commercially safe content with speed and confidence Webinar.

• Have an event, case study, gathering or interesting insight you’d like to share with the network? Drop us a note by replying to this email.

We’re a network of non-profit leaders interested in how AI is impacting our sector and driving mission. You’re receiving this email because you expressed an interest in the network. If you think this was a mistake, you can unsubscribe using the link below.