Maisy Studio
Maisy Blog

What the TechSoup AI Study Reveals About Nonprofits—and What It Misses

Maisy

Nonprofits are no longer asking whether artificial intelligence will affect their work. They are already using it.

TechSoup and Tapp Network surveyed 1,321 nonprofit professionals for The State of AI in Nonprofits 2025. The report examined how organizations use AI, what benefits they see and what prevents broader adoption.

The findings reveal a sector moving quickly—but without the structure needed to make AI reliable, secure or strategically valuable.

Adoption Does Not Equal Readiness

Many nonprofit employees now use AI for writing, research, transcription, fundraising and communications. TechSoup reports that 37.6% of respondents were already using speech-recognition tools such as Otter.ai.

But tool usage is a weak measure of organizational readiness.

A staff member using ChatGPT to draft an email does not mean the nonprofit has an AI strategy. An automated meeting summary does not mean decisions are being captured properly. Faster grant research does not guarantee that employees can locate the organization’s current program statistics, reporting requirements or approved language.

Nonprofit leaders should distinguish between three levels of adoption:

  • Individual experimentation
  • Approved departmental use
  • Governed organization-wide implementation

Most organizations begin at the first level. The danger is assuming they have reached the third.

Nonprofits Are Starting With the Easiest Uses

Content creation is attractive because it requires little preparation. Employees can open a public tool and immediately generate a donor email, social post or campaign outline.

That creates visible productivity, but it rarely solves the nonprofit’s deeper information problems.

Employees may still spend hours searching for policies. Program procedures may remain scattered across shared drives. Grant requirements may live in individual inboxes. Important decisions may disappear when experienced employees leave.

The strongest AI strategy therefore starts before the prompt. It starts by organizing the information the organization already owns.

The Real Barrier Is Not Technology

The TechSoup study highlights barriers such as limited expertise, cost, uncertainty and concern about responsible use. These are legitimate problems, but they point toward a larger issue: many nonprofits lack the internal structure needed to govern information before connecting it to AI.

Reliable nonprofit AI requires:

  • Approved source material
  • Clear ownership
  • Current policies and procedures
  • Role-based permissions
  • Human-review rules
  • Testing with realistic questions
  • A process for correcting errors
  • Defined escalation paths

Without these controls, AI can produce confident answers from outdated, conflicting or incomplete information.

AI Strategy Should Begin With Organizational Knowledge

Nonprofits should inventory where their knowledge currently exists. That includes SharePoint, Teams, OneDrive, PDFs, email, shared drives, databases and employee memory.

Then they should identify the information that creates the greatest operational risk when it is unavailable:

  • Program procedures
  • Grant requirements
  • Volunteer instructions
  • HR policies
  • Safety protocols
  • Reporting deadlines
  • Board decisions
  • Leadership history
  • Escalation procedures

Organizing this material helps the nonprofit retain institutional knowledge before experienced employees leave or responsibilities change.

A secure system such as Maisy can then help authorized employees retrieve approved information through ordinary questions, without searching across disconnected folders. The purpose is not to let AI invent organizational answers. It is to make verified answers easier to find.

The Study Is a Warning, Not Just an Adoption Report

The most important lesson from the TechSoup research is not that nonprofit AI usage is increasing. That was inevitable.

The real lesson is that adoption is moving faster than policy, training and information governance.

Nonprofits should not attempt to stop experimentation. They should identify existing use, approve appropriate tools, protect sensitive information and choose one measurable operational problem to address first.

The organizations that benefit most from AI will not necessarily be those that generate the most content or purchase the most software. They will be the ones that build reliable knowledge foundations, preserve human oversight and retain institutional knowledge as their people, programs and funding environments change.