AI adoption across membership organisations is moving quickly.
ASAE’s 2026 State of Associations research found that 87.5% of associations are already using AI for content, while 44.3% are using it with data. In the UK charity sector, 79% of organisations are now using AI in some capacity.
The ways organisations are using it are becoming more sophisticated too. AI is increasingly being used to analyse data, support monitoring and evaluation, and help people work through decisions, rather than simply drafting emails or summarising meeting notes.
And AI is arriving through another route. Increasingly, it’s being built into the CRM, analytics, communications, productivity and other platforms organisations already use.
That raises a more complicated question than whether membership organisations are ready to adopt AI. Do they understand the data environment that AI is entering?
AI readiness starts before the AI
Much of the conversation about becoming ‘AI ready’ focuses on the technology itself.
Which tools should we use? What policies do we need? How should staff be trained? Where could AI save time or improve the member experience?
All of those questions have a purpose. But underneath almost every useful application of AI sits something much less novel: data.
If an organisation wants AI to help identify engagement patterns, analyse feedback, support personalisation or surface useful insights, the information available to that system will influence what it can produce.
That does not mean giving AI access to every piece of member information an organisation holds.
Quite the opposite.
AI readiness requires organisations to understand what information they hold, where it comes from, how reliable it is, how it connects and, crucially, who or what should be allowed to access it and for what purpose.
ASAE’s guidance on AI readiness makes a similar connection between AI and underlying operational foundations, recommending that associations standardise identifiers across systems, understand data sources and flows, define responsibility for accuracy and address organisational silos.
Before asking what AI could do with your membership data, it’s worth understanding the data environment you’d be asking it to work within.
Fragmented data creates two different AI problems
The familiar phrase in conversations about AI is ‘garbage in, garbage out’.
It’s a useful principle. Inaccurate, duplicated or outdated information will inevitably limit the usefulness of analysis based upon it.
But fragmentation creates another challenge.
Membership information may be spread between a CRM, events platform, email marketing system, payment provider, community platform and assorted spreadsheets. Each source may contain perfectly accurate information while representing only one part of the member relationship.
Imagine an organisation wants to use AI to help identify members whose engagement may be declining.
Email interaction could provide a useful signal. So could event attendance, community participation, renewal history or CPD activity.
But does the AI need access to every note ever added to a member’s CRM record? Information provided for an entirely different purpose? Sensitive information about that member?
Almost certainly not.
There is another side to fragmentation too.
As different software providers add AI capabilities to their products, organisations can find themselves with multiple AI systems interacting with different parts of their data estate. This changes the question from “Can our AI see enough?”, to “Do we understand what each AI can see?”
That’s a very different kind of data-readiness problem.
Connected data should create control, not unrestricted access
For years, the case for connecting membership data has largely been an operational one.
When information is spread between multiple systems and spreadsheets, teams spend more time moving between platforms, reconciling records and manually piecing together the member journey.
Connecting that information can reduce duplication and give teams a clearer understanding of the relationship members have with their organisation.
AI potentially makes that visibility even more valuable, but perhaps not for the reason we might initially assume.
The goal shouldn’t be to give AI a 360-degree view of every member. It should be to give the organisation a 360-degree view, so it can make informed decisions about what people and technology should see.
That distinction is crucial.
If an organisation understands where its data sits, how different information relates, how sensitive it is and why it was collected, it’s in a much stronger position to decide what information an AI application genuinely needs for a particular task.
An AI tool being used to summarise anonymised survey responses has very different data requirements from one being used to identify patterns in member engagement.
Connectedness should make those decisions easier to make and govern. It should not remove the need to make them.
AI needs context, but context isn't the same as unrestricted access
There is huge potential in being able to analyse membership information more effectively.
Over time, AI could help organisations identify engagement patterns, understand feedback, recognise emerging trends or surface insights that small teams would struggle to find manually.
But useful context should be purposeful.
If an organisation wants AI to explore whether certain activities are associated with stronger retention, it might reasonably need information about membership history and participation in those activities.
That doesn’t mean it needs every piece of information the organisation holds about those individuals. This is where good data governance and good AI use begin to converge.
Membership organisations should, instead, identify what it is they are trying to understand, and the information they need to help them gain that understanding.
That’s a much healthier starting point than connecting an AI system to a database and seeing what it finds.
AI can surface a signal. A human understands the relationship
There is another limitation to even the most comprehensive dataset. Context.
A member who suddenly stops attending events might appear to be disengaging. But perhaps their role has changed. They are on parental leave. Their organisation has restructured. They have temporarily stepped back for personal reasons. Or they spoke to somebody on the membership team yesterday and explained exactly what was happening.
A system may not know that. More importantly, an AI-generated conclusion should not quietly become an organisational decision simply because it sounds convincing.
There is a significant difference between asking AI to highlight patterns that a membership team may want to investigate, and asking it to find which members are disengaged and take action.
The first supports human judgement. The second begins to replace it.
For membership organisations built around relationships, communities and individual member needs, maintaining that distinction is particularly important.
AI can help people find signals in increasingly complex information. The people responsible for members still need to interpret those signals, understand their limitations and remain accountable for what happens next.
More data isn't necessarily better data
Preparing for AI should not encourage organisations to collect information simply because it might prove useful one day.
Membership organisations still need to understand why they collect information, how accurate it is, how long it should be retained and who should have access to it.
The 2026 Charity Digital Skills Report suggests organisations are already conscious of these challenges. Skills and technical expertise remain a significant barrier to AI adoption, alongside concerns around data privacy, GDPR and security.
Those concerns shouldn’t necessarily be treated as barriers to overcome. They’re questions to answer.
The objective isn’t more ‘data + more AI’. Nor is it ‘connected data + unrestricted AI access’.
It’s appropriate, accurate, well-governed and connected data, combined with deliberate decisions about where AI can genuinely add value.
Five questions to ask before connecting AI to member data
Membership organisations don’t need a sophisticated AI strategy to begin preparing for this shift. A useful starting point is asking five relatively simple questions.
1. What are we asking the AI to do?
Start with the purpose, rather than the technology.
Summarising feedback, identifying patterns and recommending an action all involve different levels of responsibility and risk.
2. What information does it genuinely need?
Once the purpose is clear, consider the minimum appropriate information required to achieve it.
Having connected data does not mean every application should have equal access to it.
3. Do we know where that information comes from and if we can trust it?
AI does not fix inconsistent identifiers, duplicate records, poor data quality or unclear definitions.
Organisations need to understand the provenance, veracity, and limitations of the information being analysed.
4. What happens when the AI can access it?
Understand which system or model is processing the information, what happens to it afterwards, what controls exist and whether another provider or model sits behind the functionality being used.
This is becoming increasingly important as AI appears inside existing software rather than only through standalone AI products.
5. Where does human judgement and accountability sit?
Decide where AI stops and a person takes over.
Who reviews the output? When should it be questioned? What decisions should never be made without human involvement? And ultimately, who remains accountable for the action taken?
AI governance becomes much more tangible when those questions are attached to real workflows rather than left inside a policy document.
The best AI preparation may not look like AI
There is understandable pressure on membership organisations to keep pace with AI. But those best positioned to benefit from it may not necessarily be the ones adopting the greatest number of AI tools today.
They might be the ones doing much less exciting work.
Connecting systems. Improving data quality. Understanding data flows. Agreeing common definitions. Establishing ownership. Reviewing permissions. Training people to question AI outputs. Deciding where human oversight belongs.
These aren’t priorities created by AI.
They are good operational practices that already help organisations become more efficient, better informed and more consistent.
AI simply makes the consequences of getting them right, or wrong, considerably greater.
Our Head of Marketing, Kirsty Armitage, recently asked whether we’ve adopted AI faster than we’ve learned to manage it [read her article here]. One of the challenges she explored was that AI is increasingly arriving inside software organisations already use, sometimes without a conscious decision to adopt a new AI system.
For membership organisations, that makes understanding the underlying data environment even more important.
You cannot make informed decisions about what AI should access if you don’t have a clear picture of where your member information sits, how it connects, and who or what can use it.
AI will keep moving quickly. The models will improve, new capabilities will emerge and more AI will appear inside the platforms membership organisations already use.
There is no need to predict exactly what those tools will look like to start preparing for them.
The aim isn’t to make all of your membership data available to AI. It’s to understand your data well enough to decide what should be available, when, why and to whom.
That is a foundation worth building, whatever comes next.
How VeryConnect can help
VeryConnect brings member data, communications, events, payments and engagement into one connected membership platform.
By reducing the information held across disconnected systems, membership organisations can create more consistent processes and give authorised teams a clearer view of the members they support.
If you’re reviewing how well your current systems support an accessible, connected member experience, talk to the VeryConnect team.