Artificial intelligence is transforming how nonprofit organizations approach fundraising—but not all at once. Like any major technological shift, AI adoption tends to follow a predictable path, with each stage building on the capabilities and confidence developed in the previous one.
Understanding where your organization falls on this continuum can help you make smarter decisions about technology investments, staff training, data governance, and strategic planning.
One important update since most early “AI in fundraising” conversations began: the sector has moved beyond “AI = writing” and is now rapidly entering an agentic era—where AI doesn’t just help you do work, it can execute workflows inside defined guardrails (with auditability, permissions, and human oversight). This shift is now visible in major nonprofit platforms and in the broader CRM market.
Stage 1: AI-Assisted Messaging (and Productivity)
Most fundraising teams still begin their AI journey in the most intuitive place: writing and polishing communications. Drafting appeals, thank-you notes, stewardship updates, event invitations, and meeting follow-ups are natural entry points because the work is high-volume, repetitive, and relatively low-risk (if reviewed).
What’s changed recently is that Stage 1 is no longer only “copywriting.” It now commonly includes:
This is consistent with where nonprofit adoption is concentrated. In the TechSoup/Tapp Network survey (1,321 participants; Q2–Q3 2024, published as “The State of AI in Nonprofits: 2025”), 85.6% of nonprofits reported exploring or working with generative AI tools (e.g., ChatGPT and DALL-E), and “content marketing” is one of the most active current use cases.
You’re also seeing Stage 1 capabilities embedded directly into fundraising platforms. For example, Blackbaud has been rolling out generative AI experiences inside Raiser's Edge NXT—such as “Chat for Blackbaud AI,” positioned as a way to draft content and surface insights without manual searching.
Why this stage delivers immediate ROI: it shortens cycle time for communications and reduces blank-page friction, while keeping humans fully accountable for donor-facing authenticity.
Stage 2: Advanced Analytics for Your Existing Donor Base (Constituent Intelligence)
Once teams are comfortable with AI for content creation, many turn toward a more strategic application: understanding donor behavior and improving prioritization.
This stage—often called constituent intelligence—has evolved materially in the last 12–18 months. The shift is from “predictive scores” as static reports to always-on donor intelligence that:
A strong example of where the market is going: Reporting Xpress’s Constituent Intelligence (CI) Hub is positioned as an embedded donor intelligence layer that connects directly to Blackbaud Raiser's Edge or Virtuous data and turns it into proactive, automatically updating insights—flagging donors who need anniversary outreach, surfacing early warning signs of attrition, and identifying prospects who appear ready for upgrades or major gift conversations. It emphasizes actionable intelligence over static reporting, pairing segmentation and advanced RFM based analytics and lifecycle visibility with recommended next steps so teams can move from “what happened” to “what to do next.”
At the same time, this stage is still under-adopted across the sector. In the TechSoup/Tapp Network “State of AI in Nonprofits: 2025” report, only 12.8% of nonprofits reported working with predictive analytics tools—meaning Stage 2 remains a major competitive gap (and opportunity).
What’s different from “basic reporting”: basic reporting tells you what happened; donor intelligence increasingly tells you what’s likely to happen next—and what to do about it.
Stage 3: New Donor Prospect Research (AI-Augmented Prospecting)
With a stronger analytics foundation, organizations are ready to look beyond their existing donor base to identify and qualify new prospects.
The biggest recent change in Stage 3 is the emergence of AI agents for prospecting—tools that don’t just score prospects but materially reduce research time and assemble usable briefs.
For example:
Practically, Stage 3 now looks less like “researchers manually compiling profiles” and more like:
Why this stage matters: it compresses the time between “we think this person might be a fit” and “we know enough to approach them thoughtfully.”
Stage 4: Communication Personalization Based on Intent Signals (Responsive Fundraising)
This is where AI adoption moves from enhancement to transformation.
Stage 4 combines:
Crucially, “personalization” is no longer defined as mail merge. The field is explicitly calling for deeper relationship-aware messaging—e.g., referencing the donor’s demonstrated interests and the impact of their giving, not just inserting a name.
What Stage 4 increasingly includes in 2026:
This “personalize the experience” concept is now measurable in some online giving platforms. Fundraise Up, for example, publicly cites benchmarks for its AI-enabled donation experience—reporting higher click-to-donate conversion and higher average donation values compared to industry averages (as presented in their own materials).
And personalization is increasingly embedded in fundraising platforms. This is really an expansion of Stage 1, but with more autonomous actions being executed by AI.
What donors expect here (and what governance must match): Donors are increasingly open to AI if it improves relevance and stewardship, but transparency and control are now baseline expectations. In the “Donor Perceptions of AI 2025” report (1,031 US donors), 92% said it’s important that nonprofits disclose where and why AI is used and how humans remain in control.
The biggest risk here, and what is slowing adoption, is that AI-generated personalized communications will occasionally miss the mark and alienate some donors, potentially at a very high cost to the organization.
Stage 4 and Stage 5 are going to require consistent, near-perfect execution before widespread adoption happens. You will have some idea when the technology reaches this state when you stop receiving “personalized” emails from commercial vendors that you can spot as being AI-generated.
Stage 5: Autonomous AI Agent Fundraising (Digital Labor)
The frontier of AI adoption in fundraising isn’t a smarter dashboard or a faster way to draft emails inside your existing CRM. It’s autonomous fundraising capacity—AI systems that can run defined parts of moves management, at scale, within guardrails.
A clear example of this approach is Version2.ai’s Virtual Engagement Officer (VEO), which is explicitly positioned as an AI-powered autonomous fundraiser designed to “bridge the donor engagement gap” between frontline fundraisers and mass communications. The VEO is described as using traditional moves-management engagement strategies to create personalized donor journeys, re-engage lapsed donors, and deepen relationships with previously unmanaged donors—so that a gift becomes the natural outcome of an ongoing engagement process rather than a one-off campaign touch.
Importantly, this isn’t just theoretical positioning. In public rollout communications, institutions describe VEOs as expanding and personalizing donor engagement at scale, using AI-guided moves management rather than broad email blasts. It appears that organizations starting down this path are doing so with their middle-segment prospects, who are often under the radar of gift officers and may ultimately fall through the cracks. This is a “relatively” low risk medium reward place to test the waters.
Version2.ai’s recent announcements claim about $8MM in gifts successfully captured by autonomous AI agents, which is a tiny fraction of overall giving, but at least worth paying attention to.
Expect to see this trend accelerate over the next 3-5 years, unless some major public flub among early adopters scares everyone else off for a while.
What’s Next: The Agent-Native Fundraising System of Record (the real “AI-first CRM”)
Once autonomous engagement is viable, the next constraint becomes obvious: most CRMs were built as systems of record that assume humans will do the logging, updating, and hygiene work. That architecture becomes the bottleneck when you’re trying to run multiple AI roles (engagement, stewardship, planned giving, operations) alongside human fundraisers.
The “AI-first fundraising CRM” alternative to traditional CRMs is best understood as an agent-native system of record:
You can see this model emerging most clearly in the commercial CRM world among AI-native entrants.
For example, Clarify positions itself as an “autonomous CRM” that stays synced with email and calendar, with an agent (“Rep”) that handles meeting prep and follow-ups while reducing manual data entry.
Similarly, Octolane explicitly describes a “self-driving CRM” that reads Gmail and Google Calendar activity, extracts signal, and writes clean updates automatically designed to eliminate stale pipelines and manual logging.
And newer “agentic CRM” concepts like item are framing the CRM as a shared “source of truth” that gathers and enriches context across tools, then lets an assistant take actions like follow-ups or research via simple prompts (instead of complex UI).
Even Day AI frames the CRM as something that “sees what’s happening” across the pipeline and makes the next move—reinforcing the trend toward CRMs that behave more like autonomous operating systems than databases.
These new offerings are still focused on the commercial space, where workflows differ from those in nonprofits. However, expect to see similar nonprofit-centric offerings begin to appear in the next 12-24 months, probably aimed initially at the lower end of the market.
Where incumbents fit (Salesforce, HubSpot, Raiser’s Edge, etc.)
Platforms like Salesforce and HubSpot are moving aggressively into agents—but in most real-world deployments today, this still feels like agentic layers augmenting an existing CRM architecture, rather than a CRM rebuilt from first principles around ambient capture and autonomous upkeep.
These are meaningful advances—but they’re best described as the incumbent path to agentic workflows, not necessarily “AI-first CRM” in the strictest sense...
Where Does Your Team Stand?
Most fundraising teams today are between stages one and two—but the distribution is shifting quickly.
One of the most useful benchmarking snapshots comes from the TechSoup/Tapp Network “State of AI in Nonprofits: 2025” report:
That gap tells you exactly where the next competitive advantage is likely to come from: moving beyond drafting content into donor intelligence, responsive personalization, and (selective) agentic workflows.
The Bottom Line
AI will transform fundraising. The real question in 2026 is whether your organization will:
Most organizations will likely take a staged approach this year and adopt low risk high reward tools that mesh well with their existing processes. They will not want to risk falling too far behind, so they will move first into areas like automated advanced analytics and AI-powered new donor prospect research but will maintain a “wait and see” stance in terms of turning AI loose to communicate directly with their donors.