An article from guest author, Jonathan Spencer, CEO of Zetetick, a uniquie Housing Charity.
We do need to wake up about AI, rather than wrestle.
There is still a huge amount of resistance to AI across parts of the housing and supported housing sector. Some of it is sensible. The hype alone is enough to make people switch off. Every week there seems to be another person online claiming AI will either save civilisation or destroy employment entirely by next Tuesday.
Most of that noise is not especially helpful.
What interests me more is the quieter operational question underneath it all. What actually happens when organisations use these tools carefully, thoughtfully and with proper boundaries in place?
At Zetetick, we use AI in a fairly focused and deliberate way to assist us rather than overwhelm us. We have thought about it properly, taken advice and approached it pragmatically rather than ideologically. Structured workspaces, business licensing and controlled access matter because governance, GDPR and confidentiality matter enormously in supported housing. Once organisations begin understanding how enterprise AI environments actually function, many of the simplistic fears around privacy and data handling become far more manageable.
That does not mean risk disappears. AI gets things wrong. Sometimes very confidently wrong. But then people get things wrong as well. Reports get misread. Emails get misunderstood. Consultants produce weak recommendations. Managers make poor decisions under pressure. The important thing is not pretending mistakes disappear. It is building sensible verification, judgement and accountability around the technology rather than treating it either as magic or catastrophe.
In practice, the organisations using AI most effectively are usually not the ones treating it like a magic answer machine. They are the ones treating it as an ongoing conversation and learning process. In many ways, the clue is in one of the names itself: ChatGPT.
It is a discussion, not a menu.
Too many people still approach AI the way they approached Google fifteen years ago. One prompt. One answer. Quick skim. Move on. Then they conclude the technology is either brilliant or useless based on a single interaction.
That is not really where the value sits.
The value comes through iteration:
- reading outputs properly
- challenging weak reasoning
- asking for something clearer
- refining tone and structure
- requesting UK English rather than American phrasing
- removing obvious AI giveaways
- testing how communication may land with trustees, staff or commissioners
- exploring alternative approaches to difficult operational problems
Like any serious tool, AI requires skill development. Most organisations are still very early in that process, and honestly many are probably underestimating how much management time disappears each week through repetitive drafting, information overload and administrative clutter sitting around the edges of operational work.
The practical applications inside housing organisations are already significant. AI can assist with:
- analysing large documents and reports
- summarising safeguarding reviews
- supporting policy drafting
- reducing repetitive reporting work
- preparing presentations and board papers
- exploring governance or legal scenarios before formal advice is sought
- comparing lengthy guidance against existing operational practice
- helping managers structure difficult communication more thoughtfully
- reducing the amount of time spent rewriting similar material repeatedly
We have also found it surprisingly useful in areas that are harder to measure but still matter operationally. Testing how communication may emotionally land with staff groups before sending it. Simplifying dense reports without flattening meaning. Exploring whether a paper actually says what we think it says rather than what we hoped it said. Small things perhaps, although small things repeated hundreds of times across an organisation eventually become significant.
None of this replaces expertise. If anything, it exposes how important expertise still is. Weak operators using AI tend to produce weak outputs faster, while experienced operators often become calmer and more effective because some of the repetitive mental clutter reduces around them.
That distinction matters because the public conversation around AI has become strangely polarised. On one side there are endless social media figures selling “genius prompts” supposedly capable of automating success or turning people into instant experts. Most of it feels like the modern equivalent of people selling miracle business seminars years ago.
On the other side there are increasingly dramatic claims that AI will remove most jobs, destroy professions and make human contribution largely irrelevant. I do not think that is particularly serious either.
What we are actually living through is a large technological transition and, like most major transitions, it will create both opportunity and disruption. Some tasks will disappear. Others will emerge. Roles will shift. Organisations will adapt unevenly. Some people will embrace that far too quickly. Others will resist it long after the environment around them has already changed.
We have seen versions of this before. Industrial machinery changed manufacturing. Computers changed administration. The internet changed communication and commerce. None of those transitions were smooth and none arrived without anxiety, disruption or genuine social consequences. But human judgement, creativity, relationships and strategic thinking still mattered afterwards. They still do now.
In fact, I suspect genuinely human capabilities may become even more valuable as AI expands further:
- communication
- trust
- emotional intelligence
- ethical judgement
- leadership
- operational reasoning
- the ability to coordinate people coherently under pressure
Those things are not disappearing. If anything, they become more important as systems become more complex.
There is also a wider political question underneath all of this which I do not think society has properly confronted yet, because the future of AI is not simply about technology. It is also about power.
A very small number of global technology corporations now hold extraordinary influence over systems increasingly shaping communication, information, labour, behaviour and decision-making across society. Governments are still trying to understand technologies they are simultaneously attempting to regulate, while large technology firms are moving at enormous speed backed by levels of capital and influence that would have seemed extraordinary not that long ago.
People sometimes dismiss concerns around “tech bros” as exaggerated rhetoric. I think that is probably naïve. When a small number of unelected corporations and individuals begin shaping the informational and technological infrastructure underneath daily life, labour and communication, that becomes a serious political and social issue whether people like the language surrounding it or not.
None of this means AI should be rejected. But it does mean organisations should approach it consciously rather than blindly.
And honestly, avoiding AI entirely is becoming its own form of organisational risk, particularly in sectors already carrying enormous administrative and operational pressure. Supported housing organisations are dealing with increasing safeguarding complexity, heavier reporting requirements, commissioner scrutiny, staffing pressure and operational overload. A reasonably skilled operational manager using AI carefully can already save several hours a week by reducing repetitive administrative and cognitive work sitting around the edges of the role.
That does not remove the need for good people. It creates more space for them to think, communicate and lead more effectively. And supported housing probably needs that thinking space badly.
If anyone in housing, care or supported living wants a sensible and grounded conversation about AI, governance, operational systems or how organisations can use these tools carefully without losing judgement or values, feel free to get in touch. The sector needs more honest conversations about this and fewer extremes.
Sources and further reading:
National Housing Federation, How Housing Associations are Adapting to AI (2026). Found that many housing associations still report low AI confidence and limited organisational readiness.
Housing Technology Magazine, Artificial Intelligence: The Dawn of a New Era in UK Social Housing (2023). Useful examples around operational efficiency, analytics and housing systems integration.
UK Information Commissioner’s Office (ICO), Guidance on AI and Data Protection. Important practical guidance around GDPR, governance and responsible AI use within organisations.
