Mango Signal App Overview
Mango Signal turns workforce signals into action. It computes a continuous engagement health score across the company, every segment, and every manager team, surfaces top themes and risk segments, recommends opinionated next steps, drives a close-the-loop “you said, we did” flow back to employees, and measures the impact of every completed action.
What is Mango Signal?
Mango Signal is the Workforce Intelligence layer for MangoApps. Where Surveys collects feedback and Report Agent answers ad-hoc data questions, Mango Signal interprets cross-signal context, prioritizes action for managers, drives change, and proves improvement. It is not a new collection tool; it sits on top of the apps you already pay for.
The core problem it solves: most organizations collect employee feedback but rarely act on it. Surveys produce dashboards. Comments pile up. Managers don’t follow through. Employees never see what changed. Trust drops, participation falls, and the cycle repeats. Mango Signal closes that loop.
Who is it for?
- CHROs and VP People — for the rolling engagement health score, top themes, and impact narrative.
- Managers and team leads — for team-scoped insight, AI-recommended actions, and the close-the-loop flow.
- People Analytics teams — for the cross-signal data layer and CSV/XLSX exports.
- Internal Communications leaders — for the close-the-loop posts to News Feed and Broadcast.
- VP Operations and plant/store ops directors — for risk monitoring across teams, locations, and segments.
Primary surfaces
Dashboard (everyone)
The default landing page. Shows the latest engagement health score, top themes from the last 30 days, your open actions, and (for managers) risk segments with AI-recommended actions you can convert into real action items in one click.
Signals (manager+)
Read-only browse of every cross-app signal Mango Signal has ingested. Filter by source app, signal type, segment, or time window. Useful for drilling from a risk segment down to the underlying source events (e.g., “what survey responses contributed to the engineering segment’s drop?”).
Actions (manager+)
The action workflow. Create, edit, complete, or dismiss actions through a drawer-style form. Every action can link back to originating signals and forward to a measured Impact. Filter by status, assignee, or “assigned to me.”
Impact (manager+)
The proof tab. For every completed action, see the baseline engagement score, the post-action score, and the delta. Aggregate counts at the top: how many actions improved sentiment, how many didn’t, the average delta.
Analytics (admin)
Cross-signal aggregation: engagement score over time, signal volume, theme breakdown, risk segments at the configured threshold, action / impact summary. Export to CSV or XLSX.
Settings (admin)
Toggle the three core feature areas (engagement scoring, risk alerts, close-the-loop), tune the score recompute window, set the risk alert threshold, and load/destroy demo data for previewing.
Mango Signal AI assistant (Ask AI)
An opinionated AI assistant focused on culture, engagement, and risk. It answers questions like “what’s our engagement score this month,” “which teams are sliding,” “recommend actions for the warehouse team’s recognition drop,” and (with confirmation) creates or completes actions on your behalf. Risky writes always require explicit user approval before persisting.
How signals flow in
Mango Signal does not generate signals — upstream apps do. v1 ships writers for the most common sources:
- Surveys — sentiment scores from per-response AI analysis (when the Surveys app is wired with the duck-typed interface)
- Recognitions — every Recognition post emits a
recognitionsignal scoped to the recipient - EPMS (Performance) — review submissions, goal status changes, and 360 feedback emit
performancesignals scoped to the subject - News Feed — engagement events (likes, comments, acks) emit
engagementsignals scoped to the actor - Shift Feedback — every submission emits a
sentimentsignal scoped to the user (or the location for anonymous submissions)
Additional apps in the strategy plan (Tasks, Workspace, Timesheets, Attendance, Leave, Safety Hub, Inspections, Service Desk) are wired in subsequent phases.
A one-time backfill job sweeps the prior 90 days of upstream data so day-one of Mango Signal looks populated. Re-running the backfill is idempotent — signals dedup on (source_record_type, source_record_id).
How the engagement score is computed
The score is a weighted aggregate over the last 30 days (configurable):
| Signal type | Weight | Notes |
|---|---|---|
| Sentiment | 0.40 | Normalized -1..1 → 0..1 (sentiment is the only type stored on -1..1) |
| Recognition | 0.20 | Already 0..1 at ingestion |
| Performance | 0.15 | Normalized from the rating scale (1..5 or 1..10) at ingestion |
| Engagement | 0.15 | Already 0..1 at ingestion |
| Theme | 0.10 | Counted, not scored |
The job (MangoSignalRecomputeScoresJob) writes a new MangoSignal::EngagementScore row per scope (company / segment / team / location / manager) on every run. History is append-only — trends, deltas, and pre/post-action comparisons are queryable without re-aggregating raw signals.
Settings reference
| Setting | Type | Default | Effect |
|---|---|---|---|
engagement_score_enabled |
boolean | true | Gates MangoSignalRecomputeScoresJob and the dashboard score widget |
risk_alerts_enabled |
boolean | true | Gates MangoSignalMailer#risk_alert |
close_loop_enabled |
boolean | true | Gates the close-the-loop fan-out (News Feed / Broadcast) |
agent_enabled |
boolean | true | Gates the Ask AI agent |
score_recompute_window_days |
number | 30 | Look-back window the score job uses |
risk_alert_threshold |
number | 0.4 | 0..1; segments below this are flagged at risk |
What’s coming (roadmap)
- Phase 2 — Operational Intelligence: attrition / retention risk, burnout risk, coverage / scheduling risk, manager effectiveness scoring, workforce capacity, safety / compliance trend intelligence, talent readiness.
- Phase 3 — Predictive and Prescriptive: forecasts, recommended interventions with ROI tracking, executive workforce planning views, benchmark comparisons.
See docs/strategy/MANGO_SIGNAL_PLAN.md for the full strategy plan.
FAQ
Is Mango Signal a separate app from Surveys? Yes. Mango Signal interprets cross-signal context across many apps; Surveys is one of those apps.
Does it expose individual sentiment? No. The agent and dashboards always aggregate to team, segment, or location scope.
What if I don’t have Surveys / Recognitions / EPMS? Mango Signal works with whatever upstream apps you have. The more apps emit signals, the richer the cross-signal interpretation.
Can I disable the AI assistant? Yes — flip agent_enabled off in Settings.
Where do Mango Signal exports come from? The Analytics tab includes CSV and XLSX export buttons. The data is the raw signal table for the configured window, with all subject and segment dimensions preserved.
Last updated: 2026-05-06. App version: 1.0.0.