The rise of AI agents
While productivity tools such as Microsoft Copilot currently dominate AI adoption across social housing, the research also identifies interest in AI agents.
Seventeen respondents, 25%, reported using AI agents, showing that this technology is already present within part of the respondent base, although it remains less common than general AI tools.
Unlike general-purpose AI tools that assist with writing, research or analysis, AI agents are typically designed to carry out specific tasks or support defined workflows. The examples identified through the survey demonstrate a range of emerging applications, from information retrieval through to operational support and automation.
"Examples suggest some organisations are beginning to explore how agents can move AI beyond general productivity support"
Respondents described examples including:
- Policy and knowledge retrieval assistants
- Repairs support tools
- Project management assistants
- Contact centre support
- Workflow automation
- Customer service applications
- Operational task support
One organisation reported developing two AI agents capable of retrieving policies and repairs schedule-of-rates information on demand. Another referenced agents being used for project management support, while several respondents highlighted AI-assisted automation within customer service and operational processes.
The qualitative responses generally describe agents in the context of experimentation, defined tasks and individual operational applications rather than widespread organisation-level deployment.
This is reflected elsewhere in the research. Only seven respondents reported having a formal AI strategy and only three reported measuring AI impact consistently. These findings suggest agent adoption is developing within a wider environment where many organisations are still building the governance, skills and measurement frameworks needed to support AI at scale.
While adoption remains relatively limited, the examples suggest some organisations are beginning to explore how agents can move AI beyond general productivity support and into defined housing-service workflows and task-based operational applications.
