From Fragmentation to Foundation – Closing the Infrastructure Gap

AI adoption is accelerating across the nonprofit sector, but the systems supporting it are not always keeping pace.

This challenge sits at the heart of the second gap identified in the recent US research from the Blackbaud Institute, Bridging the AI Effectiveness Gap1: the Infrastructure Gap. Central to this research are four common barriers that can limit organisations’ ability to translate AI adoption into meaningful organisational value: Effectiveness, Infrastructure, Data-Readiness, and Transparency.

This article is the second in a four-part series exploring each of these gaps through an ANZ fundraising lens. In the previous article, we discussed why organisations often struggle to translate AI adoption into effective use. This article looks at the next challenge: what happens when AI adoption outpaces the systems designed to support it.

The ANZ findings point in a similar direction, with the challenge appearing even more pronounced. Insights from the Status of ANZ Fundraising 20262 suggest that while technology use is widespread, relatively few organisations report having fully integrated systems to support it. Likewise, in the US, organisations often struggle to build the foundations needed to support growing AI use, creating risks around governance, consistency, and scalability.

 

What is the Infrastructure Gap?

The Infrastructure Gap is the gap between widespread AI use and the organisational systems needed to support it safely and consistently. In the US research, AI adoption is often individual rather than systemic, with only 50% of organisations using paid or enterprise AI tools and 24% relying exclusively on free tools. The findings point to a disconnect between growing AI use and the shared systems, governance, and approved tools needed to support it at scale.

Unmanaged AI tools, including free tools used without clear oversight, can create additional risks around security, privacy, and data handling, particularly when sensitive information is involved. Without shared tools, processes, and oversight, organisations can find it difficult to move from AI as an individual effort to a scalable organisational capability.

 

Do we see the same challenge in the ANZ?

Yes, and the ANZ evidence is particularly strong.

The Status of ANZ Fundraising 2026 report shows that technology use is high but fragmented, with only 5% of organisations saying their technology stack is well integrated and 17% saying it is not integrated at all. The findings also highlight training to use technology fully (66%), improved data management (65%), more modern technology, including AI (58%), and integrated solutions (57%) as some of the technology priorities valued most by nonprofits. In addition, technical support for integrating AI with existing systems (59%) is the most commonly requested form of AI support. Together, these findings suggest that infrastructure, rather than willingness to adopt AI, has become one of the biggest barriers to creating greater value from it.

 

Closing the Infrastructure Gap: What ANZ Fundraisers Should Do Next

The Infrastructure Gap is not solved by adopting more technology. It is solved by creating the foundations that allow technology to be used consistently, securely, and at scale. For ANZ fundraising organisations, that means focusing on a few practical actions that reduce fragmentation, strengthen integration, and help teams use AI with greater confidence and consistency.

  • Move from personal AI habits to an approved AI toolset: For most organisations, that means adopting a small number of approved tools, setting clear boundaries around data use, and making sure someone is responsible for reviewing issues or exceptions when they arise.
  • Prioritise integration over tool sprawl: In ANZ fundraising teams, the bigger win is often a cleaner supporter workflow and fewer disconnected systems, rather than another AI app. Fragmentation is already one of the sector’s biggest challenges, and when systems do not work well together, even the most useful AI tools can increase friction rather than improve efficiency.
  • Create lightweight operating rules, not heavyweight bureaucracy: A practical AI policy should cover approved use cases, prohibited uses, data handling, review points, and who signs off exceptions. The goal is not to create additional layers of process, but to provide more clarity. Well-defined operating rules can also help organisations navigate broader governance obligations and expectations under the Australian Charities and Not-for-profits Commission (ACNC) framework and Charities Services New Zealand guidance.
  • Bring fundraising ethics into AI infrastructure decisions: AI infrastructure decisions should not be viewed solely as technology decisions. Consider how tools, suppliers, data handling practices, and governance arrangements support responsible fundraising. This aligns with principles reflected in the Fundraising Institute Australia (FIA) Code, which highlights the importance of transparency, ethical behaviour, accountability, and the responsible stewardship of supporter information.
  • Avoid unmanaged free tools when supporter data is involved: Before introducing new AI tools, understand how external providers handle supporter information, integrate with existing systems, and support organisational requirements. Vendor review should be treated as a core fundraising governance task and should also be underpinned by the FIA Code, as well as guidance from the Office of the Australian Information Commissioner (OAIC) and the New Zealand Privacy Commissioner.

 

Closing thoughts

The Infrastructure Gap highlights a simple but important reality: using AI at scale requires more than access to AI tools.

Both reports point to the same conclusion: the challenge is no longer technology availability or organisations’ willingness to adopt AI. It is whether they have the infrastructure needed to support it consistently, securely, and at scale. Fragmented systems make it harder to share data, standardise processes, and apply AI effectively across teams.

For ANZ fundraising organisations, the opportunity is to move beyond isolated AI use and build the foundations needed to support it across teams, workflows, and fundraising activities. Focusing on integration, clear ownership, proportionate governance, and responsible data handling helps turn AI from a collection of individual productivity gains into a capability that can deliver value across the organisation.

Next in the series: The Data-Readiness Gap

If the Effectiveness Gap is about creating value from AI, and the Infrastructure Gap is about creating the foundations that allow that value to scale, the next challenge is ensuring those foundations are built on reliable data. In the next article, we’ll explore the Data-Readiness Gap and look at why data quality, accessibility, and consistency remain essential for successful AI adoption across fundraising organisations.

Research Notes

  • 1Blackbaud Institute, Bridging the AI Effectiveness Gap, 2026
    The Bridging the AI Effectiveness Gap report is based on two surveys conducted in March 2026 in the United States by the Blackbaud Institute and Edge Research. The study included 1,389 social impact professionals and 1,034 donors. As the findings reflect US organisations and donor attitudes, they should not be treated as directly comparable to ANZ. However, the underlying pattern is highly transferable: in both the US and ANZ, AI adoption is moving faster than organisational readiness, and the strongest results are achieved by organisations that combine AI with clearer goals, better data, stronger governance, and deliberate trust-building.
  • 2 Blackbaud, Status of ANZ Fundraising, 2026
    The Status of ANZ Fundraising 2026 report is based on a January and February 2026 survey of 246 ANZ participants, and 60% of the sample represents organisations in the top 10% of ANZ charities by revenue. That means the ANZ findings are directionally strong, but smaller charities may face tighter capacity constraints than the averages suggest; recommendations therefore need to be proportionate for small and mid-sized organisations as well as larger ones.

Further Reading