🗞️ AI for Non Profits Network: Weekly Briefing 12/16

The weekly digest from a network of non-profits interested in AI. What's in this week's Briefing: American Heart Association's AI Playbook; What Vertical AI Bundles Mean for Fundraising; + resources

In The Briefing this week:

1. 👀 What’s Caught our Eye: American Heart Association’s “Proof Over Hype” AI Playbook

2. 💭 Thought for the Week: The Stack Is Collapsing: What Vertical AI Bundles Mean for Nonprofit Fundraising

3. 🌟 Interesting News

4. 🌐 From Across the Network

1) 👀 What Caught Our Eye: American Heart Association’s “Proof Over Hype” AI Playbook

I’ve been reading CEO Nancy Brown’s recent field notes from HLTH 2025 on how the American Heart Association is bringing AI into a high-trust system, and it’s one of the clearest articulations I’ve seen of what mature AI adoption actually looks like at scale.

Here’s the context: AHA is a $1.2 billion nonprofit with global reach, 35 million volunteers and supporters, and a century-long reputation built on rigorous science. They’re not just experimenting with AI for efficiency gains - they’re figuring out how to maintain trust while technology reshapes cardiovascular care delivery.

What they’ve come up with is an independent AI Assessment Lab, built in partnership with Dandelion Health, that evaluates cardiovascular AI tools against real-world clinical data (EMRs, ECGs, echocardiograms) before health systems deploy them. They’re testing for clinical effectiveness, ethical integrity, and operational integration in what CEO Nancy Brown calls an “increasingly crowded, often confusing” AI marketplace.

This isn’t just vendor vetting. It’s infrastructure for trust. AHA is creating a quality standard for an entire sector, using their scientific credibility to help smaller health systems “no matter their size or geography” make safer decisions about which AI tools actually work.

Why this matters beyond health:

Most nonprofits won’t build assessment labs. But the pattern transfers: if you’re adopting AI in youth services, mental health support, housing assistance, or education - anywhere trust is your currency - you need an evaluation lane before deployment.

At smaller scale, that might look like: a risk assessment rubric, structured pilot testing with real users and clear success metrics, documented kill-switch criteria, or partnerships with academic evaluators who can provide independent validation.

The discipline matters more than the sophistication. AHA’s Assessment Lab is the Rung 4 version of what every nonprofit needs at Rung 2: proof that something works before you bet your reputation on it.

Brown’s underlying message is trust isn’t built by moving fast. It’s built by moving carefully, with evidence, and with the humility to test before you scale.

Are you deploying AI in a health context - share your project here so we can spotlight it to the network

2) 💭 Thought of the Week: The Stack Is Collapsing: What Vertical AI Bundles Mean for Nonprofit Fundraising

In the network, we often look at other sectors to see what might make its way across to the non-profit space. This week we came across a post from Jason Shuman at Primary Ventures that’s been circulating in tech circles, and it’s making me think differently about what’s coming for nonprofit fundraising infrastructure.

His thesis is simple: the era of disconnected point tools is ending. Across every major industry, companies are replacing their patchwork of sales and marketing software with what he calls “vertical AI bundles” - integrated systems that own the entire revenue lifecycle from first signal to expansion, built specifically for one sector.

His argument is simple: “Selling revenue is easier than selling software because the numbers don’t lie.”

When a product can credibly demonstrate it helped generate measurable outcomes - more revenue, higher conversion, better retention - you’re not debating feature lists. You’re buying results.

The Nonprofit Translation

Nonprofits don’t run a “revenue lifecycle” in the classic sense. But we do run something equally measurable and equally system-driven: the supporter lifecycle - from discovery through first gift, stewardship, upgrade, long-term retention, and advocacy.

And our stack is collapsing too.

Most organizations I talk to are managing a familiar patchwork:

• CRM plus email platform

• donation forms plus payment processing

• event tools

• peer-to-peer platforms

• spreadsheets for segmentation

• communication calendars living in people’s heads

• and stewardship that depends entirely on over-stretched staff remembering to follow up.

It works - until it doesn’t.

The gaps show up as missed follow-ups after donations, inconsistent supporter communications, generic asks that don’t match context, low recurring conversion, weak reactivation of lapsed donors, and teams drowning in administration rather than relationship-building.

What a Vertical Bundle Actually Means

Shuman’s insight is that AI companies aren’t winning by building better versions of existing tools. They’re winning by building integrated systems that orchestrate the entire workflow - and proving they drive outcomes.

He describes how these bundles typically develop: companies start with a “wedge” - one painful, high-urgency problem like automated follow-ups or proposal generation. That wedge gets adoption. Then the breakout companies layer in the rest of the stack until the product becomes what he calls a “Revenue Operating System” for that vertical.

For nonprofits, I’m convinced the equivalent is emerging: a Virtual Engagement Officer plus Fundraising Operating System that orchestrates the supporter journey end-to-end.

Here’s what that could look like, translated from Shuman’s framework:

Supporter Discovery (not lead generation) — Identifying who is warming up, who is ready for an ask, who should be nurtured versus contacted directly.

Fundraiser Coaching (not sales coaching) — Turning every staff member into a stronger fundraiser with outreach drafts, objection handling, meeting prep, call notes that become next steps.

Stewardship Automation (not follow-ups) — Immediate thank-you flows, impact updates tied to giving, event follow-ups, lapsed donor reactivation. This is where supporter relationships break down in most organizations.

Appeals and Grant Builder (not proposal creation) — Generating sponsorship decks, grant narratives, campaign pages, appeal sequences in your organization’s voice and workflow.

Ask and Checkout Engine (not CPQ) — Optimizing suggested amounts, one-time versus recurring prompts, matching gift nudges, fee coverage, Gift Aid logic, mobile-first conversion.

Campaign Workflow OS — Orchestrating content calendars, approvals, segmentation, task assignment, compliance guardrails, and handoffs between humans and AI agents.

Engagement Brain (not revenue brain) — Choosing next-best action: who to call today, who gets stewardship, who should see a different ask, who is at risk of lapsing, who is ready to upgrade.

Donor Data and Intent Signals — A live signals layer connecting CRM, email behavior, web activity, event attendance, donation history. Not to surveil supporters, but to stop being blind.

Retention and Growth Engine — Driving recurring conversion and upgrades, reactivating lapsed supporters, building major donor pathways, renewing corporate partnerships, activating peer-to-peer loops.

Why This Matters Now

The pattern he’s seeing in the corporate world - where AI makes it possible to collapse 7-10 disconnected tools into one integrated system that proves outcomes - that same pattern is coming for nonprofit fundraising. The question is whether we’ll see products built specifically for our sector’s needs, or whether we’ll be stuck adapting corporate sales tools that don’t understand stewardship, trust-building, or long-term donor relationships.

The lessons for non-profits go back to similar themes we’ve highlighted across the network:

• Choose a desired outcome and clear metric that can be improved.

• Start with a ‘wedge’ or small pilot on a specific pain point.

• Prove value there, and then scale across the rest of the vertical.

What We’re Looking Out For

If vertical AI bundles really are the next wave, nonprofits will need our own version - one built around stewardship, trust, and long-term relationships rather than aggressive sales motion.

But there are also risks. The temptation to automate relationship-building entirely. The pressure to optimize for short-term conversion over long-term trust. The gap between organizations that can afford integrated systems and those stuck with free or low-cost fragments.

The question isn’t whether this is coming. It’s whether we shape it intentionally, with our sector’s values embedded from the start.

After reading about vertical AI bundles collapsing the traditional sales and marketing stack, I’m curious where your organization is with fundraising, CRM and engagement tools?

3) 🌟 Interesting News

Artificial intelligence trends for 2026 (Charity Digital)

OpenAI, Anthropic, and Others Create Foundation for Standardizing AI Agents (Gizmodo)

How a non-profit research lab produced France’s hottest new AI start-up (Science Business)

4) 🌐 From Across the Network

• We’re launching an AI Capability Workshop to develop how your non profit implements AI. Get in touch to be involved.

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

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.