About & data use

The application collects de-identified stakeholder perspectives about the principles and trade-offs involved in employee benefits decision-making. It complements qualitative stakeholder interviews by providing breadth and structured evidence.

The tool intentionally avoids collecting participant identity and sensitive personal data. Participation is anonymous by default and there is no participant profiling, scoring or ranking.

• Random response identifier
• Broad stakeholder group
• Structured comparison responses
• Statement responses
• Point allocations
• Optional general comments
• Instrument version
• Completion timestamp and completion duration
• In Pilot / Calibration Mode only, optional interview participation type (app only / interview + app / prefer not to say); app responses are not linked to individual interview transcripts

• Names
• Email addresses
• Employee IDs
• Manager or reporting information
• Precise job titles
• Business unit, legal entity, office location or precise country
• Compensation information
• Individual benefit enrolment or utilisation information
• Demographic or other sensitive personal information
• Health or disability information
• Confidential business data

File uploads are not supported and participants are not asked about individual cases.

The application does not make employment, eligibility, compensation, benefits or other HR decisions about individuals.

Results inform organisational research and strategy discussions. They do not constitute promises, employee votes or automatic policy decisions. Stakeholder perspectives are one evidence layer alongside qualitative context, feasibility and organisational judgement.

Generative AI is not required to calculate the research results or determine the benefits philosophy, and is not used to interpret individual participants.

The application performs deterministic mathematical calculations required for pairwise weighting, aggregation and reporting: counts, percentages, distributions, medians, quantiles, logarithmic least-squares estimation of relative weights from incomplete comparison sets, and reporting-threshold suppression.

Any future generative AI analysis would be treated as a separate feature requiring appropriate governance approval.

• Stakeholder-group comparisons are suppressed below a configurable minimum of 5 completed responses
• Percentages are always shown with the underlying respondent count
• Distributions are shown rather than averages alone
• "I don't have enough information to form a view" is recorded separately rather than assigned to a midpoint
• Responses are stored against a random identifier only
• The application never converts a result into a recommendation or an organisational decision