Here’s a problem: deeply understanding our members is hard.

It's easy to make assumptions about our members' needs and offer the same services and support we've always done. But if we want to grow in numbers and influence, it's worth digging deeper into what truly motivates members to join, stay and get active.

The challenge is that traditional methods for understanding members – surveys, focus groups, phone banks – are either too shallow to capture nuance or too time-consuming to do at scale. This is the tension unions face: how do you understand the diverse needs of thousands of members when you barely have time to process the feedback you're already getting?

This was the question at the heart of a recent webinar run by Unions 21 and Agenda, where Unions 21’s Nick Scott and Agenda’s Vic Barlow were joined by union staff from around the world to explore if AI might offer new routes to deepen our understanding of members’ needs.

Rather than just talking about the theory, Nick ran a live experiment to demonstrate some ways AI can collect qualitative data at scale, process it and give clues as to what people are thinking and why. Watch the webinar below:

First, participants completed a short voice-powered survey about their experiences with AI – speaking their answers naturally rather than typing them. The responses appeared on screen anonymously in real-time.

Within minutes, a series of AI tools had transcribed everyone's spoken responses, identified key themes, performed sentiment analysis, and even created detailed audience personas based on how people thought about AI. The kind of work that would normally take a team days or weeks happened in real-time during the session.

When participants had entered their locations in various formats – full country names, abbreviations, cities – AI managed to convert this messy data into a properly formatted Excel spreadsheet with standardised countries and regions. Data cleaning, which often takes so long it gets skipped entirely, happened instantly.

Then, from the free-text responses about AI opportunities and concerns, recurring themes like data security, efficiency gains and time savings were identified. AI spotted patterns in how people with high confidence about AI still had specific worries, and how those with concerns still saw potential benefits. Nuanced insights that might be missed when you're drowning in raw data.

Then it created audience personas – profiles like "cautious organiser" and "pragmatic experimenter" – complete with their priorities and concerns.

So what does this mean in practice for union work?

Victoria Barlow, co-director of Agenda and strategic communications consultant for unions, puts it bluntly: members are not a homogeneous group. They have varied experiences, problems and expectations from their union. When we fail to understand these nuances, we end up primarily addressing the needs of the largest or most easily reachable groups, potentially alienating others.

She shares examples from her work where deeper understanding made all the difference. At Balpa, the British airline pilots' association, it turned out that members were deeply frustrated about car parking. Not a standard industrial issue in this sector, but it significantly impacted their daily lives and their relationship with the union. At the Association of Teachers and Lecturers, really listening to teachers' detailed feedback about workload led to creating a calculator that empowered them to push back against excessive demands.

Unions need to take time to truly understand what members care about. But here's the problem: that requires collecting rich qualitative data from as many members as possible, then making sense of it.

Here are some examples of where and how AI might be able to support:

As Vic emphasised, this isn't about replacing human relationships or organising conversations. It's about making those conversations more powerful by understanding member needs better. It's about enabling face-to-face organising to be more targeted and effective, not substituting it with automated systems.

But the webinar didn't shy away from the uncomfortable questions this technology raises. How comfortable are members with voice-powered surveys? Might the data collection method itself change what people feel willing to share? How do we ensure AI-processed data gets checked by humans before we make decisions based on it?

The answers aren't simple, but some principles are clear:

Human oversight is essential. AI might spot patterns, but it can also miss context or make errors. Someone needs to review what the AI produces, especially before integrating it into membership systems or making strategic decisions.

Quick data isn't necessarily good data. Just because you can collect and analyse responses rapidly doesn't mean they're representative of your actual membership. You still need to think about who's responding and who isn't.

The tool shapes the data. Different ways of asking questions get different kinds of answers. A voice survey might feel more natural to some members and intimidating to others. We need to think about how our methods might be influencing what we hear.

Participants in the webinar raised practical questions too. What about data security? Where does the data actually go when you use these tools? These are all important and complicated questions… and exactly the kinds of questions Unions 21 tackles in its AI training courses for union staff.

Ultimately, the question isn't whether AI can help us understand members better… used effectively, there are many ways this technology could support unions. The question is how to use it responsibly. That means thinking carefully about:

Training for staff. Most union staff have received no training on AI. Using these tools effectively requires understanding both what they can do and what can go wrong.

Clear policies. Who can use which tools? What data can be processed? How do we check the AI's work? Organisations like the TUC have developed model policies, but every union needs to think through these questions for their own context.

Keeping humans at the centre. AI should enhance organisers' ability to connect with members, not replace those connections. The goal is stronger relationships, not efficiency at the expense of solidarity.

Vic's closing point made it clear: this is about making unions stronger by understanding members better. It's about turning the technology of our time toward the same goal unions have always pursued – supporting workers and building power.

AI transparency statement: Claude.ai generated a first draft of this article, based on a transcript of the webinar and using set style guides. The article was then edited by a human.