03

03

03

Predictive Staffing

Predictive Staffing

Predictive Staffing

The Predictive Staffing module helps you manage active schedules and proactively anticipates staffing demands across multiple departments or clusters. Predictive Staffing provides census forecasts, sicks calls, and makes it easy to direct float pool resources where they are needed most, ensuring smooth day of operations and streamlining workflows between charge RNs and staffing offices.

The Predictive Staffing module helps you manage active schedules and proactively anticipates staffing demands across multiple departments or clusters. Predictive Staffing provides census forecasts, sicks calls, and makes it easy to direct float pool resources where they are needed most, ensuring smooth day of operations and streamlining workflows between charge RNs and staffing offices.

The Predictive Staffing module helps you manage active schedules and proactively anticipates staffing demands across multiple departments or clusters. Predictive Staffing provides census forecasts, sicks calls, and makes it easy to direct float pool resources where they are needed most, ensuring smooth day of operations and streamlining workflows between charge RNs and staffing offices.

Company

Product

Role
Duration
Platform

Aya Healthcare

Predictive Staffing

Senior Product Designer (IC)

2024 - Present

Enterprise Web (B2B Saas)

Impact

Impact

Impact

  • Enabled healthcare organizations to proactively identify staffing gaps before they occurred

  • Improved staffing efficiency and reduced reactive scheduling decisions

  • Contributed to enterprise adoption as part of Aya’s Workforce AI product suite-

Problem

Problem

Problem

Healthcare staffing teams often operate in a reactive mode — responding to shortages only after they impact operations.


Existing staffing processes:

  • Relied on historical data rather than forward-looking insights

  • Made it difficult to identify risk until gaps became urgent

  • Required manual coordination across multiple systems and teams


The challenge:
Design a predictive staffing experience that surfaces risk early, enables proactive action, and integrates naturally into existing staffing workflows — without creating alert fatigue or mistrust in AI recommendations.

My Role & Scope

My Role & Scope

My Role & Scope

I led product design for Predictive Staffing from early concept through enterprise launch, working closely with:

  • Product Management

  • Engineering

  • Data Science

  • Operational stakeholders

  • Sales and Executive teams


Responsibilities included:

  • Shaping product experience strategy in a highly ambiguous domain

  • Translating predictive models into actionable staffing insights

  • Designing workflows that supported proactive decision-making

  • Prototyping concepts used in executive reviews and enterprise sales

Users & Context

Users & Context

Users & Context

Primary users:

  • Staffing coordinators

  • Workforce planners

  • Operational leaders


These users needed to:

  • Identify potential staffing gaps before they became urgent

  • Prioritize limited staffing resources effectively

  • Take action quickly when risk was identified


Key constraint:
Staffing decisions were time-sensitive and high-impact — false positives or unclear recommendations could erode trust and slow adoption.

Key Constraints &

Design Challenges

Key Constraints & Design
Challenges

Key Constraints & Design
Challenges

  • Prediction confidence: Users needed to understand how likely a staffing gap was to occur

  • Alert fatigue: Too many warnings would reduce effectiveness

  • Workflow fit: The product had to integrate into fast-paced staffing operations

  • Trust in automation: Recommendations needed to feel supportive, not prescriptive


These constraints guided both interaction and visual design decisions.

Design Approach

Design Approach

Design Approach

Rather than treating predictions as static alerts, I focused on:

  • Risk-based prioritization over binary warnings

  • Action-oriented insights instead of raw predictions

  • Contextual explanations that built trust without slowing workflows


Close collaboration with Data Science ensured the UI reflected both model behavior and real-world staffing realities.

Key Design Decisions

Key Design Decisions

Key Design Decisions

1. Surface risk as a spectrum, not a binary state


Instead of “staffed / not staffed,” I designed a graded risk model that:

  • Highlighted varying levels of staffing vulnerability

  • Helped users prioritize attention and resources

  • Reduced overreaction to low-confidence predictions

1. Surface risk as a spectrum, not a binary state


Instead of “staffed / not staffed,” I designed a graded risk model that:

  • Highlighted varying levels of staffing vulnerability

  • Helped users prioritize attention and resources

  • Reduced overreaction to low-confidence predictions

2. Design recommendations around existing workflows


Staffing teams already had established processes. Rather than forcing new behaviors, I:

  • Embedded predictions into familiar views

  • Designed actions that aligned with current decision patterns

  • Reduced cognitive load during time-sensitive tasks

2. Design recommendations around existing workflows


Staffing teams already had established processes. Rather than forcing new behaviors, I:

  • Embedded predictions into familiar views

  • Designed actions that aligned with current decision patterns

  • Reduced cognitive load during time-sensitive tasks

3. Balance automation with human control


To avoid the perception of “AI replacing judgment,” the experience:

  • Presented recommendations with clear rationale

  • Allowed users to override or adjust actions

  • Emphasized decision support over automation

3. Balance automation with human control


To avoid the perception of “AI replacing judgment,” the experience:

  • Presented recommendations with clear rationale

  • Allowed users to override or adjust actions

  • Emphasized decision support over automation

Predictive Staffing System Overview

Predictive Staffing

System Overview

A modular system for staffing decisions at different time horizons and organizational levels
Designed Predictive Staffing as a connected set of modules, each optimized for a specific decision type rather than a single overloaded experience

Short-term, operational decisions:

  • Weekly Schedule

  • Lookahead


Structural, organizational decisions:

  • Unit Staffing

  • Central Staffing

A modular system for staffing decisions at different time horizons and organizational levels
Designed Predictive Staffing as a connected set of modules, each optimized for a specific decision type rather than a single overloaded experience

Short-term, operational decisions:

  • Weekly Schedule

  • Lookahead


Structural, organizational decisions:

  • Unit Staffing

  • Central Staffing

Weekly Schedule Overview

Weekly Schedule

Overview

Designed for rapid, in-week staffing decisions
Prioritized clarity and scanability to support time-sensitive adjustments without disrupting existing workflows.

Designed for rapid, in-week staffing decisions
Prioritized clarity and scanability to support time-sensitive adjustments without disrupting existing workflows.

Weekly Schedule Detail

Weekly Schedule Detail

Contextual detail without workflow interruption
Surfaced explanatory signals only when needed to preserve speed and focus.

Contextual detail without workflow interruption
Surfaced explanatory signals only when needed to preserve speed and focus.

Lookahead Forecast View

Lookahead Forecast View

Early warning for future staffing risk
Designed Lookahead to shift teams from reactive staffing to proactive planning weeks in advance.

Early warning for future staffing risk
Designed Lookahead to shift teams from reactive staffing to proactive planning weeks in advance.

Lookahead Detail

Lookahead Detail

Built trust through interpretability
Emphasized directional trends and confidence rather than point predictions to avoid false certainty.

Built trust through interpretability
Emphasized directional trends and confidence rather than point predictions to avoid false certainty.

Lookahead UXR

Lookahead UXR

Built trust through interpretability
Emphasized directional trends and confidence rather than point predictions to avoid false certainty.

Built trust through interpretability
Emphasized directional trends and confidence rather than point predictions to avoid false certainty.

Built trust through interpretability
Emphasized directional trends and confidence rather than point predictions to avoid false certainty.

Built trust through interpretability
Emphasized directional trends and confidence rather than point predictions to avoid false certainty.

Built trust through interpretability
Emphasized directional trends and confidence rather than point predictions to avoid false certainty.

Built trust through interpretability
Emphasized directional trends and confidence rather than point predictions to avoid false certainty.

Built trust through interpretability
Emphasized directional trends and confidence rather than point predictions to avoid false certainty.

Built trust through interpretability
Emphasized directional trends and confidence rather than point predictions to avoid false certainty.

Unit Staffing View - EDIT

Unit Staffing View

- EDIT

EDIT
EDIT

EDIT
EDIT

Unit Staffing View

Unit Staffing View

Staffing risk contextualized at the unit level
Enabled leaders to identify systemic staffing issues rather than isolated incidents.

Staffing risk contextualized at the unit level
Enabled leaders to identify systemic staffing issues rather than isolated incidents.

Unit Staffing View 2

Schedule Template

Prototype

EDIT
EDIT

EDIT
EDIT

Unit Staffing View 3

Schedule Template

Prototype

EDIT
EDIT

EDIT
EDIT

Central Staffing View

Schedule Template

Prototype

Centralized visibility for coordinated staffing decisions
Designed for leaders managing staffing across multiple units and locations.

Centralized visibility for coordinated staffing decisions
Designed for leaders managing staffing across multiple units and locations.

Central Staffing View 2

Schedule Template

Prototype

EDIT
EDIT

EDIT
EDIT

Central Staffing View 3

Schedule Template

Prototype

EDIT
EDIT

EDIT
EDIT

Central Staffing View 4

Schedule Template

Prototype

EDIT
EDIT

EDIT
EDIT

Solution

Solution

Solution

The final Predictive Staffing experience included:

  • Risk-based staffing forecasts highlighting potential gaps

  • Prioritized recommendations aligned with operational urgency

  • Clear explanations to support trust in predictive insights

  • Seamless integration into existing staffing workflows


The product helped teams shift from reactive firefighting to proactive staffing management.

Product Experience Overview

Product Experience
Overview

Together, these modules support staffing decisions across time horizons and organizational levels — from in-week adjustments to long-range, cross-unit planning — while maintaining a consistent mental model and trust in predictive insights.

Together, these modules support staffing decisions across time horizons and organizational levels — from in-week adjustments to long-range, cross-unit planning — while maintaining a consistent mental model and trust in predictive insights.

Outcome & Impact

Outcome & Impact

Outcome & Impact

  • Predictive Staffing became a core feature within Aya’s Workforce AI platform

  • Operational teams identified staffing risks earlier and acted more decisively

  • Reduced last-minute staffing interventions and improved planning confidence

  • Strengthened Aya’s enterprise AI value proposition in sales and RFP discussions

Reflection

Reflection

Reflection

This project reinforced several design principles:

  • Predictive tools must prioritize interpretability over precision

  • Trust is built through transparency and control, not automation

  • Aligning AI insights with existing workflows accelerates adoption


If revisiting this work, I would explore:

  • Personalization of risk thresholds by organization

  • Deeper feedback loops to improve prediction accuracy over time

  • Expanded cross-product integrations within the Workforce AI ecosystem

Let's work

together.

Let's work

together.