Why a lack of AI understanding is holding the public sector back

Growing interest, limited confidence

Our research found that many housing organisations were actively exploring AI, with around 60% researching how it could support their work. However, only a small percentage had reached advanced stages of adoption. This pattern is mirrored across the wider public sector.

Teams are curious about AI. Leaders recognise its potential to improve efficiency, reduce admin, and support better decision-making. But curiosity doesn’t always translate into confidence. Without clear guidance, many organisations remain stuck in “exploration mode”.

In practice, this means AI is often used in an informal, unstructured way (if at all) rather than as part of a coordinated strategy.

The knowledge gap problem

One of the most striking findings from our report was that 87% of housing associations reported low levels of AI knowledge. When applied more broadly, this highlights a widespread skills gap across public services.

This isn’t about staff lacking ability or motivation. It’s about access to the right training, time, and support. Many public sector teams are already stretched, juggling funding pressures and rising demand. Finding space to learn about emerging technologies can feel impossible.

Without strong AI literacy, organisations risk:

• Underusing valuable tools
• Misunderstanding risks
• Losing confidence in innovation
• Falling behind more digitally mature peers

Concerns around risk, ethics, and governance

Another major barrier is uncertainty around safety and responsibility. This is particularly important in the public sector, where organisations handle sensitive personal data and make decisions that affect people’s lives. Concerns about data protection, bias, and transparency are valid, but without the right frameworks, they can lead to inaction.

Many organisations told us they had no formal AI strategy, acceptable use policy, or governance structure in place. Without these foundations, leaders are understandably cautious about scaling AI use.

The result? Valuable opportunities are delayed or missed altogether.

The missed opportunity

Despite these challenges, our research also highlighted strong optimism. Most respondents believed AI could improve efficiency, productivity, and service delivery.

Imagine what this means for the wider public sector:

• Faster case handling in social services
• Smarter demand forecasting in healthcare
• More responsive citizen services
• Reduced administrative burden for frontline staff

AI has the potential to free up time for what matters most: supporting people and communities. But these benefits only become reality when organisations move beyond experimentation and build real capability.

From uncertainty to confidence

So, how can public sector organisations bridge the AI understanding gap?

From our work across housing and beyond, a few principles stand out:

Start with people, not technology
Investing in AI literacy helps staff feel confident, capable, and involved in change.

Build strong governance early
Clear policies and ethical frameworks create trust and reduce risk.

Focus on practical use cases
Small, low-risk pilots help demonstrate value and build momentum.

Partner with specialists
External expertise can accelerate learning and reduce trial-and-error.

Most importantly, organisations need space to ask questions, test ideas, and learn without fear of getting it wrong.

Looking ahead

AI will continue to play a growing role in public services. The question is not whether organisations will adopt it, but how well prepared they will be when they do.

Phoenix’s housing research shows that lack of understanding is currently the biggest barrier. But it also shows something else: a sector that wants to learn, adapt, and improve.

Continue the conversation: join our Copilot webinars

Understanding AI is one thing. Using it safely and effectively at scale is another. To help public sector organisations move from exploration to impact, we’re hosting two practical Copilot webinars.

How to scale Copilot safely focuses on the foundations you need to get right, from licensing and security to governance and ROI. It’s ideal for organisations running pilots or preparing for wider rollout.

How to maximise your Copilot investment in 2026 looks at adoption, usage, and long-term value. If Copilot is already in your organisation but impact feels inconsistent, this session will help you turn access into results.

Both sessions are designed for public sector teams looking to build confidence, control risk, and unlock real value from AI.