Methodology
Survey design and fieldwork
This index analyses the environment for AI adoption in the public sector across ten markets: Brazil, the United States, France, Germany, the United Kingdom, South Africa, Saudi Arabia (KSA), India, Japan, and Singapore.
The findings are based on an online survey of 3,335 public sector workers, with approximately 300 respondents per market. The sample drew from a wide range of public sector organisations and roles, with respondents from local or regional government entities (54%), national government or national government entities (36%), and other public sector organisations (10%). It also reflected a mix of seniority levels, including office or operational staff in non-managerial roles (42%), managers or supervisors (29%), frontline or service delivery workers (13%), and senior leaders or executives (11%). Fieldwork was conducted between 15 November and 5 December 2025.
Respondents were recruited via established online research panels, where participants opt in to take part in surveys in return for a small financial incentive. Panel providers contacted respondents on our behalf and administered the survey. Where relevant, surveys were translated into national or widely spoken languages to ensure accessibility and comprehension.
Data quality and validation
To maximise the reliability and robustness of the data, a series of standard quality checks were applied prior to analysis. These included:
- Excluding respondents who completed the survey unusually quickly or slowly;
- Excluding respondents who engaged in straight-lining behaviour (e.g. selecting the same response option across multiple items);
- Excluding respondents who failed embedded attention checks;
- Excluding responses that were duplicated or perfectly matched those of another respondent;
- Reviewing and excluding incoherent open-text responses, including those that appeared automated or bot-generated.
All fieldwork was conducted in line with the rules and guidance of the British Polling Council (BPC) and the Market Research Society (MRS). Public First is a member of both organisations.
Index framework and construction
The Public Sector AI Experience Index assesses the conditions for AI adoption in the public sector across five conceptual dimensions. These dimensions are grounded directly in survey questions testing attitudes, behaviours, resources, and organisational context.
The five index dimensions are:
- Enthusiasm – whether public sector workers feel positive, optimistic, and motivated about AI in the workplace;
- Empowerment – whether workers feel they have clear permission, leadership backing, and governance clarity to use AI at work;
- Enablement – whether workers have access to appropriate AI tools, resources, and technical support;
- Embedding – whether AI is integrated into everyday workflows, systems, and organisational processes;
Education – whether workers feel trained, knowledgeable, and confident in how to use AI.
Each dimension is constructed from a set of survey questions grouped based on conceptual relevance and empirical coherence.
All questions were recoded so that higher values consistently indicate more mature or supportive conditions for AI adoption. Barrier questions were reverse-coded, governance balance questions were scored to reward perceived balance, and related items were combined using simple averages or counts to capture breadth of experience or support.
For each respondent, bucket scores were calculated as the average of normalised (0–1) scores for all items within that bucket. The overall Index was then calculated as the unweighted average of the five bucket scores and rescaled to a 0–100 index, with country scores derived by averaging respondent-level results.