Empatalk maps your Communication DNA across 22 factors organised into 5 categories. Each factor is a signal we capture from your survey answers and feed into the AI generators that produce your guideline, comparisons, and rewrites.

This page is the canonical reference for what we measure. If a factor name shows up in your profile, your DNA radar, or a comparison verdict, you can come here to understand what it represents.

The survey is the only place these values get set. We never derive them from anything else — no log scraping, no behavioural inference, no third-party data enrichment.

_Who you are. Sets the broad context the AI uses to calibrate everything else._

•  Name (text) — How you want to be addressed.

•  Roles (multi-select) — What you do day to day (Engineer, Manager, Founder, Designer, Operator, Researcher, Marketer, Salesperson, Student, Teacher, Therapist).

•  Communication Goal (text) — What you want to get better at (clearer writing, better feedback, managing conflict, connecting with team, understanding others).

•  Age range (enum) — Broad bracket: 18–24, 25–34, 35–44, 45–54, 55+.

•  Country (enum) — Where you are. Used for cultural context, not advertising.

•  Education path (enum) — Self-taught, on-the-job, undergrad, postgrad.

•  Neurodivergence (multi-select, optional) — ADHD, autism, dyslexia, anxiety, HSP, other. You can also choose "Prefer not to say". Stored with `visibility: private` by default; never shown publicly.

•  Interests (multi-select) — The metaphor reservoir the AI draws from when rewriting your messages. ~30 options including skiing, cooking, AI, photography, woodworking, chess, gardening, sci-fi, surfing, poetry.

_How other people should communicate with you. These shape what shows up in your "How to work with me" guideline and what the Assistant tells someone messaging you._

•  Directness (numeric, 0–100) — How blunt you want messages addressed to you. 0 = soften everything, hedge with care; 100 = lead with the headline, skip the preamble.

•  Detail Level (enum) — How much context you want: Less, Balanced, More.

•  Feedback Preference (enum) — How you receive critique best: Private (DM), In-the-flow (during the work), Written (async + reviewable).

_How you naturally communicate. Used when the AI rewrites a draft you authored — preserves your voice instead of homogenising it._

•  Sarcasm Tolerance (enum) — Whether sarcasm lands for you when used by others. Drives whether the Assistant suggests rewrites that match your dry register, or filters it out for recipients who would miss the irony.

•  Irony Tolerance (enum) — Same dimension as sarcasm but for indirect statements. Some communicators use irony as a precision tool; others find it confusing.

•  Assumption Sensitivity (enum) — How much explicit verification you need before agreeing on shared context. High sensitivity = "let me restate what I think we just agreed".

•  Ambiguity Tolerance (enum) — Whether you do well with open-ended directives or need precise scope.

•  Sender Nuances (multi-select, up to 3) — Compressed v2 chip representation of the four sender-style factors above for survey UX velocity. Options include "Sarcasm lands", "Irony lands", "Avoid assumptions", "Comfortable with ambiguity", "Need precision", "Detailed by default".

_Environmental signals — your norms, frictions, and energisers. Help the AI flag mismatches before they happen._

•  Communication Challenges (multi-select, up to 3) — Patterns you find hard: Saying no, Giving hard feedback, Speaking up in groups, Async-only contexts, Cross-cultural nuance, Conflict.

•  Boundaries (multi-select, up to 3) — Communication norms you want held: Async-first, No Slack after hours, One-topic-per-message, Video for hard talks, Written before sync, Quick acks expected.

•  Triggers (multi-select, up to 3) — Patterns that activate negative reactions in you: Vague feedback, Passive-aggression, Info overload, Public criticism, Slow replies, Unclear expectations.

•  Glimmers (multi-select, up to 3) — Micro-moments that energise you: Someone really listens, Real talk, Quick acks, Voice notes, Visual diagrams, Deep dives.

_How the AI should help you. Direct controls over the Assistant's behaviour, not derived from anything else._

•  AI Assistance Mode (enum) — How aggressive the Assistant gets when you paste a draft: Rewrite my words (full rewrite), Suggest options (multiple alternatives), Just check tone (commentary only, no edits).

•  Default Tone (enum) — What baseline register the Assistant defaults to when you don't specify a recipient: Direct, Diplomatic, Warm, Concise.

---

Total: 22 factors. These power the DNA radar on `/u/<your-slug>`, your guideline, the compatibility-comparison engine, and every Assistant rewrite. You can edit any answer from your account settings — changes propagate to the next AI generation immediately.