Your platform tracks employee sentiment and adoption. What employee data are you actually collecting, and how do I avoid this becoming a surveillance tool?

The full executive answer

The honest answer is: far less than you might fear, by deliberate design. The Culture, Adoption & Employee Sentiment module (M38) works at team level, not individual level — its records are team, tool, adoption percentage, a one-to-five sentiment score, and change load. There are no per-employee behavioural profiles, no keystroke or activity monitoring, and the Talent Intelligence module (M16) models capability gaps at the role level — hire, upskill, contract decisions — not named-person assessments.

On top of that, three platform-wide protections apply. Any free text that reaches an AI model passes through an injection-neutralisation and PII-redaction layer, so a comment naming a colleague is scrubbed before a model ever sees it. All data is tenant-isolated at the database layer with row-level security. And every access is written to an append-only audit log, so your data protection officer can verify who saw what.

Current versus roadmap: team-level aggregation is how it works today, which I would argue is the right ceiling for a boardroom tool. If a customer ever asks for individual-level analytics, our position is that it needs their own DPIA and works-council sign-off first — we will not quietly enable it.