Hear how your team
really feels.
Short, anonymous check-ins right inside Slack and Teams — turned into clear, shareable signal about engagement, belonging and wellbeing.
Pulse surveys measure the organization. Match, meeting and program surveys measure individual coffee chats.
engagement dimensions
research-backed questions
pre-built templates
anonymous by default
Most tools tell you engagement dropped.
We do something about it.
Measure
Anonymous pulses land in Slack and Teams. Scores, trends and focus areas compute every cycle.
Act
Belonging low in Engineering? Launch a cross-team connection program against exactly that group — same platform, same audiences, no export.
Re-measure
Next cycle scores the same dimension for the same segment. You see whether it moved.
Standalone survey tools give you step one. CoffeePals gives you all three.
View connection programsFrom question to insight
in four steps.
Pick a template — or build your own
A guided six-step wizard: details → audience → schedule → messages → privacy → review. Start from one of 7 templates or assemble your own set from the 131-question library.
Pulses arrive in Slack or Teams
Questions are delivered one at a time — answer the first and the next appears, so a 4-question pulse never looks like a form. One automatic reminder fires partway through the window.
People answer anonymously
One tap for Likert, rating and eNPS; a text box for open text. Score low and a conditional follow-up can ask what is driving it, in the moment.
Insights appear automatically
When the window closes, the platform computes dimension scores, a composite, eNPS, segment breakdowns and cycle-over-cycle trends — then writes a plain-English headline, key movements, risk callouts and a ranked top-3 focus areas.
AI-written summaries require AI features, included in Max. Without them, all scores, trends, segments and reports still work.
The engagement survey
people actually answer.
Nobody drops out of a survey they finish in under a minute.
1–5 questions, one at a time
Default 4. Delivered as a native card in DM, answered with a tap. Answer the first and the next appears — it never looks like a form.
Smart rotation
The engine picks the dimension asked longest ago, then the least-recently-asked question inside it. The anti-repeat window defaults to 8 weeks on weekly programs, 4 on biweekly and monthly.
See it before you send it
Rotation preview shows exactly which questions go out over the next cycles. eNPS injects every 4 weeks by default — alongside the regular questions or replacing a slot, your call.
Anonymity you can explain
to your most sceptical engineer.
Everyone has been burned by a survey tool that leaked identity. Here is exactly how this one cannot.
Anonymous by default
New programs are created anonymous, with a confidentiality threshold of 5.
Nothing shows until enough people answer
No dimension score, composite, eNPS or text theme is displayed unless enough distinct people responded — 5 by default, configurable from 3 to 20. The count is of distinct respondents, not response rows, so one person answering six questions can never clear the threshold alone.
Settings can only get stricter
Once a program is live, the anonymity mode is locked, the threshold can be raised but never lowered, and two-way dialogue can be switched off but never back on.
No hidden link back to the person
For anonymous programs without dialogue, responses store no pointer to the user at all — not even a hashed one, because a deterministic hash would let anyone with database access cluster a person’s answers across cycles.
- Scores aren't computed below 3 responses
- Segments need 5+ participants
- Cross-tabs capped at 2 attributes, each cell gated independently
- Verbatim quotes stripped below 10 distinct respondents
- Composite requires 3+ dimensions, all independently clearing the threshold
- eNPS promoter and detractor splits hidden below the threshold
Reply to an anonymous comment
without ever learning who wrote it.
“We heard you” only means something if the person who spoke up hears back.
- An anonymous comment lands in the program’s inbox.
- The admin clicks reply and writes a message.
- The participant gets a DM in Slack or Teams, quoting their original question and answer for context.
- They reply in chat. The thread continues, with unread counts on both sides.
- The admin never sees a name — routing runs through an opaque, encrypted handle.
The guardrail
A dialogue cannot be started at all on a comment whose question group has fewer respondents than the confidentiality threshold. Contacting “the” respondent would narrow identity, so the product does not allow it.
Under a minute to answer.
Zero minutes to analyse.
Leadership-ready reports
Executive Summary
A high-level engagement snapshot for leadership.
Trend Report
Cycle-over-cycle analysis with significance flags.
Every section toggles on or off, and each carries built-in “what it measures” and “how to read it” guidance written into the report — so a non-specialist can present it without coaching.
Three CSV exports: dimension scores, eNPS and participation. Closed cycles only, so an in-progress cycle never produces a misleading partial row.
Scores every cycle
Dimension scores on a normalized 1–5 scale with favorable, neutral and unfavorable splits, a composite, and eNPS from −100 to +100.
Segments with benchmarks
Scores by department, location, tenure, manager or anything synced from your HRIS. Every segment row carries the org benchmark and its delta, so you see instantly who is above and below.
Trends you can trust
A rolling three-cycle window per dimension, classified as improving, declining, stable or volatile. Variance thresholds scale per metric, so eNPS’s wider range is not flagged volatile on every normal wobble.
Focus areas
A ranked top 3 to act on, each with a written explanation and its top theme, scored from decline magnitude, unfavorability, response volume and segment variance. When everything is stable or improving, it reframes as strengths to maintain rather than manufacturing problems.
Theme extraction, cycle summaries and focus-area explanations require AI features, included in Max.
We didn’t make
these questions up.
Every question in the library carries a research note explaining why it is asked and what it predicts. Build your own dimensions alongside these 14.
Belonging
Inclusion, acceptance, and connection within the organization
Growth & Development
Opportunities for learning, career advancement, skill development
Manager Effectiveness
Quality of management support, feedback, and communication
Recognition
How well contributions are acknowledged and appreciated
Autonomy
Independence and empowerment in daily work
Alignment
Understanding of and connection to mission and goals
Wellbeing
Work-life balance, stress levels, overall wellness
Collaboration
Teamwork, cooperation, cross-functional relationships
Communication
Effectiveness and transparency of organizational communication
eNPS
Employee Net Promoter Score — likelihood to recommend the organization
Psychological Safety
Whether people feel safe to speak up, take risks, and be vulnerable without blame
Trust in Leadership
Confidence in senior leaders' competence, integrity, and care
Workload & Burnout
Workload sustainability, exhaustion, burnout risk
Diversity, Equity & Inclusion
Fair treatment, equal opportunity, inclusive culture across all backgrounds
Questions, answered.
Start listening in Slack or Teams.
Included in Max, alongside advanced matching, connection analytics and HRIS user import.