Access does not equal adoption
One of the clearest findings from the research is that providing access to AI tools does not automatically lead to widespread adoption. While many organisations have made AI available to staff, far fewer have been able to embed it consistently across teams, services and day-to-day working practices.
Access to AI tools was widespread among respondents. Thirty-nine respondents, 59%, reported that all or most staff have access to AI tools provided by their organisation. A further 17 said access was available to some teams, while seven said access was restricted to a small number of staff and one respondent reported no access at all.
However, the picture changes significantly when usage is considered. Only nine respondents reported that more than half of staff use AI tools at least once a week. Eighteen estimated regular usage at between 25% and 50% of staff, while 12 said adoption was between 10% and 25%. Ten respondents believed fewer than 10% of colleagues used AI regularly. Perhaps most tellingly, 17 respondents said they did not know how many staff were using AI tools at least once a week.
This gap between availability and usage suggests that organisations are encountering a challenge familiar to many technology adoption programmes: providing access is relatively straightforward, but changing behaviours is considerably harder.
The survey responses reinforce this conclusion. Twenty-six respondents said AI use depends largely on individual teams, making it the most common description of adoption across the dataset. Fifteen reported that AI use happens informally without much guidance, while another 15 described it as encouraged but inconsistent. Only nine respondents said AI use was actively encouraged and supported across the organisation.
"Many organisations remain reliant on enthusiastic individuals, local champions and early adopters. In practice, this often means that AI adoption develops unevenly."
Confidence in AI
These findings indicate that many organisations remain reliant on enthusiastic individuals, local champions and early adopters. In practice, this often means that AI adoption develops unevenly. Some teams experiment extensively and identify valuable use cases, while others make little use of the tools available to them.
Training and workforce confidence appear to be significant factors. Twenty-four respondents reported that no formal support had been provided for AI adoption. The most commonly cited forms of support were optional training and AI champions or communities of practice, each selected by 21 respondents. Written guidance was selected by 13 respondents, while only four reported mandatory training.
Confidence levels were also mixed. Half of respondents were neutral when asked how confident staff feel using AI effectively. Eighteen, 27%, said colleagues were not very confident and a further two said they were not confident at all. Just 13 respondents described staff as fairly confident.
Training and support provided for AI use
No formal support
AI champions or community if practice
Optional training
Written guidance only
External training
Mandatory training
Leadership and sponsorship
Leadership engagement may also play an important role. Only eight respondents described AI adoption as executive-led and actively championed. By contrast, 19 reported limited executive involvement and another 13 said there was no visible executive sponsorship. While AI activity is clearly taking place across the respondent base, the findings show that active executive sponsorship remains uncommon among respondents.
Although the number is small, all three respondents who described AI as an organisation-wide strategic priority also reported active executive sponsorship. Among respondents at earlier stages of maturity, involvement was more frequently described as limited, delegated or not visible. This is an association within the survey rather than evidence that sponsorship alone causes higher maturity.
For many respondents, the challenge extends beyond obtaining AI tools to building the confidence, support and working practices needed to use them consistently. The findings suggest that social housing's next challenge is not providing access to AI, but creating the conditions that enable colleagues to use it effectively and at scale.
Housemark believes organisations will be better placed to convert access into meaningful adoption when technology is supported by training, leadership, appropriate governance and practical use cases. The challenge is no longer simply making AI available. It is making AI useful, trusted and embedded within everyday work.
How actively does your executive team sponsor AI adoption?
Limited executive involvement
Supported but delegated
No visible executive sponsorship
Executive-led and actively championed
Don’t know
Housemark expert insight:
Why access alone doesn't drive adoption
One of the most common misconceptions surrounding AI is that adoption naturally follows access. In reality, organisations often discover that making tools available is only the beginning of the journey.
The findings indicate that access is most likely to translate into consistent use when colleagues receive training, guidance and encouragement to apply AI to practical aspects of their work. Clear leadership, visible support and examples of successful use can all help build confidence and reduce uncertainty.
Technology enables adoption, but people determine whether it succeeds. Organisations that invest time in developing skills, sharing good practice and supporting colleagues through change are more likely to realise value from the tools they already have available.
John Wickenden,
Research Manager, Housemark
Spotlight:
The rise of the AI-powered analyst
One of the most distinctive themes emerging from the research is the growing role of AI within data, insight and performance teams.
Thirty-five respondents, 53%, reported using AI for data analysis, while 27, 41%, said they use it for reporting and dashboards and 24, 36%, for coding or technical work. Across the survey, respondents repeatedly referenced SQL development, Power BI reporting, dashboard creation, scripting, workflow automation and technical problem-solving as valuable AI use cases.
Several respondents highlighted the role of AI in supporting SQL and DAX development, creating dashboards, automating reporting tasks and improving access to organisational information. Others described using AI to troubleshoot technical issues, write reporting code and accelerate analytical work that would otherwise require significant manual effort.
Examples identified across the survey included:
- SQL query creation and optimisation
- DAX development
- Power BI dashboard creation
- Performance reporting
- Regulatory reporting
- Workflow automation
- Coding support
- Data analysis and insight generation
This type of AI-assisted development is sometimes described as “vibe coding”, where generative AI is used to accelerate coding, development and technical problem-solving. The survey does not measure respondents’ underlying technical expertise, but the examples suggest AI is becoming a practical development assistant for some analysts and practitioners.
While still developing, it may prove to be one of the most significant ways AI changes day-to-day work within social housing. The responses suggest AI is enabling a wider range of practitioners to experiment with automation, reporting, dashboard creation and technical problem-solving.
For many respondents, AI is no longer just a productivity tool. It is increasingly becoming a practical development and analytical assistant, helping colleagues create reports, build dashboards, automate repetitive tasks and generate insights from organisational data.
Key takeaway
For Housemark, one of the most interesting themes emerging from the research is the growing role of AI as a practical assistant for analysts, insight teams and practitioners. The examples point towards faster experimentation with reporting, dashboards, scripts and workflows, alongside a continuing need for appropriate governance, oversight and quality control.