Workforce Intelligence

Workforce Intelligence

1. What it is

Workforce Intelligence is the admin surface where you configure two systems that work together: MangoScore, which evaluates employee reliability from attendance, cancellations, manager ratings, and consistency signals, and Recommendation AI, which uses configurable weights to rank candidates when the scheduling engine assigns shifts.

• Enablement: requires the Workforce Intelligence permission, which is admin-only — it is not inherited by Manager roles.
• What it is not: this is not a marketplace app you enable per-tenant. It is a built-in admin surface that appears under Admin → Workforce Intelligence in the sidebar for any user with the permission.

2. Standing it up

1. Grant the permission. Go to Admin → Roles & Permissions and grant Workforce Intelligence → View (or Manage) to the roles that should access this surface. Only admin-level roles receive it — Manager roles do not inherit it.

2. Open the dashboard. Navigate to Admin → Workforce Intelligence. The dashboard shows four stat tiles: Average MangoScore, Total Employees, Scored Employees, and whether Recommendation AI is Active or Setup Needed. Until the first weekly calculation runs, the Average tile shows a dash.

3. Configure MangoScore weights. Click Scoring & Logic → MangoScore Settings. Set the component weights (attendance, cancellation, manager rating, consistency) — they must sum to 1.0. Review the tier thresholds and adjust if your defaults need to change.

4. Configure Recommendation Weights. Click the Recommendation Weights tab. Set the four allocation weights (availability, preference, fairness, role) — they must sum to 100%. Optionally configure contracted-hours enforcement.

5. Wait for the first calculation. MangoScore recalculates automatically every Sunday at 02:00 UTC. Employees need at least 5 assigned shifts in the last 30 days to receive a score — anyone below that threshold is placed in the Unranked (Low Activity) tier. You can also trigger an immediate recalculation from the MangoScore settings (saving changed weights or thresholds automatically enqueues one).

3. How it fits together

MangoScore is a per-employee reliability number (0.0–1.0) computed from four components, each with a configurable weight:

• Attendance — proportion of assigned shifts the employee showed up for (default weight: 50%)
• Cancellation — proportion of unexcused late cancellations or no-shows, inverted so lower is better (default: 30%)
• Manager rating — average of 1–5 ratings entered by managers, normalized to 0–1 (default: 20%)
• Consistency — punctuality streak length and shift-swap reliability, equally weighted; a successfully covered swap is neutral, only uncovered drops penalize (default: 0%, opt-in)

The score places each employee into one of four tiers based on configurable thresholds:

Tier	Default threshold	Meaning
Top Tier	≥ 0.90	Highest reliability; eligible for early shift access
Reliable	≥ 0.75	Solid performer
Watchlist	≥ 0.00	Needs attention; managers are notified on entry
Unranked (Low Activity)	—	Fewer than 5 shifts in the last 30 days

All four tier names are customizable per-tenant.

Recommendation AI is a separate weighting system that the scheduling engine’s auto-assign uses to rank candidates for a shift. Its four weights (availability, preference, fairness, role) must sum to 100%. A separate contracted-hours control sits outside the sum — it scales the penalty applied as an employee approaches their contracted-hours target, and the enforcement mode decides whether the cap is a hard limit, a soft penalty, or off.

How they connect: MangoScore feeds into the scheduling engine’s candidate ranking alongside the Recommendation AI weights. Top Tier employees can also claim shifts during the early-access window before the shift becomes generally available.

4. Running it

Viewing employee scores

Go to Admin → MangoScore (linked from the Workforce Intelligence dashboard). The index lists all employees with their current score, tier, attendance rate, cancellation rate, and manager rating. Filter by tier, location, role, or name. Export the current view as CSV.

Rating an employee

On the MangoScore detail page for an employee, enter a 1–5 manager rating. The score recalculates immediately. You cannot rate yourself — the system blocks self-ratings.

Triggering a recalculation

From the MangoScore index, use Recalculate All to enqueue a background recalculation for every employee in the business. The weekly job does this automatically each Sunday. Saving changed MangoScore weights or thresholds also triggers a recalculation.

Adjusting recommendation weights

Open the Recommendation Weights tab under Scoring & Logic. Drag the sliders for availability, preference, fairness, and role — the stacked bar updates live to show the balance. The total must reach exactly 100% before the form can be saved.

Configuring contracted-hours enforcement

On the same Recommendation Weights page, set the contracted-hours weight multiplier and choose an enforcement mode:

• Strict (default) — employees who would exceed their contracted-hours target are filtered out of auto-assign entirely
• Soft only — they stay in the pool but receive a score penalty
• Off — contracted hours are ignored in auto-assign

5. Settings

MangoScore settings

Navigate to Admin → Workforce Intelligence → Scoring & Logic → MangoScore Settings.

Setting	Default	What it changes
System Name	MangoScore™	Display name used throughout the application for the scoring system
Attendance Weight	0.50	How much attendance contributes to the score
Cancellation Weight	0.30	How much cancellation behavior contributes
Manager Rating Weight	0.20	How much manager ratings contribute
Consistency Weight	0.00	How much punctuality streak and swap reliability contribute; 0 disables this component
Top Tier Threshold	0.90	Minimum score to reach Top Tier
Reliable Threshold	0.75	Minimum score to reach Reliable
Watchlist Threshold	0.00	Minimum score for Watchlist (below Reliable)
Tier Names (×4)	Top Tier / Reliable / Watchlist / Unranked (Low Activity)	Custom display names for each tier
Early Access Window	24 hours	How many hours before general release Top Tier employees can claim shifts
Weekly Digest	Off	When on, managers receive a weekly notification summarizing their team’s scores

Recommendation weights

Navigate to Admin → Workforce Intelligence → Scoring & Logic → Recommendation Weights.

Setting	Default	What it changes
Availability Weight	30%	Weight given to employee availability for a shift
Preference Weight	25%	Weight given to employee shift preferences
Fairness Weight	25%	Weight given to even distribution of shifts
Role Weight	20%	Weight given to role-match for the shift
Contracted Hours Weight	1.0×	Multiplier on the soft penalty as utilization approaches 100%
Contracted Hours Enforcement	Strict	Whether the contracted-hours cap is a hard limit, soft penalty, or off

6. More help

• Workforce Intelligence FAQ — specific setup and operating questions (coming soon)
• Ask AI — the assistant answers questions about MangoScore and Recommendation AI from these articles.