Gathering and analysing market survey results using AI
From: Government Actuary's Department (GAD) | Published 2 October 2025
GAD supported the Ministry of Housing, Communities and Local Government (MHCLG) in analysing a market survey focused on high-rise buildings. The work combined structured responses with free-text analysis to produce clear, actionable evidence for policy.
Unique survey design that turned text into evidence
The survey went beyond tick-box questions and captured free-text responses to surface nuance. Using artificial intelligence, GAD converted unstructured comments into structured fields, reduced duplication, and standardised entries for consistent analysis.
This approach allowed teams to quantify sentiment and themes at scale while keeping important context. The result: faster synthesis of findings that can be tested, compared, and communicated.
Interactive dashboard built for decision-makers
Working with MHCLG, GAD delivered an interactive dashboard that lets users interrogate the data without specialist support. It highlights trends, context, and movement over time in a way that is clear and operational.
- Drill into specific topics and themes
- Identify geographic variation and local patterns
- Visualise changes over time for quick comparison
The dashboard shortens briefing cycles and supports faster responses to policy questions.
Cross-specialist collaboration
The project brought together actuaries, policy officials, economists, and delivery specialists. This mix of skills ensured the work met diverse policy and analytical needs while maintaining technical rigour.
Findings informed MHCLG's approach to assessing options and value for money. The use of generative AI and modern data visualisation delivered meaningful efficiency savings and supported evidence-based decisions.
Enhanced understanding
Actuary Jonathan Day said: "Through this collaboration, the project team now better understands the dynamics of the market of interest. The project showcased the effectiveness of using multidiscipline teams from across government to overcome complex problems."
This clarity helps target future interventions more effectively and supports MHCLG's objectives.
What this means for government teams
- Capture nuance: include free-text fields where appropriate and plan for analysis from the start (taxonomies, coding guides, and validation steps).
- Standardise early: clean and harmonise entities (locations, organisations, building types) to make comparisons defensible.
- Keep humans in the loop: combine AI classification with sampling, review, and clear audit trails.
- Build for questions, not data: design dashboards around the policy decisions they must support.
- Address ethics and governance upfront: document data sources, model choices, and known limitations. See the OECD AI Principles for high-level guidance.
Further learning
If you are upskilling policy or analysis teams on AI, see practical course paths by role at Complete AI Training.
Your membership also unlocks:
