Building a Resilient Healthcare Workforce with AI Staff Rosters

Mike Peterson is the Associate Clinical Manager for Radiology at Whanganui Hospital in New Zealand. With a background entirely in clinical work, Mike manages the day-to-day running of the department including rostering, leave compliance, committees, and numerous projects.

Before implementing RosterLab, the previous manager spent at least 40 hours (a full week's work) creating each 12-week roster using Excel spreadsheets. Mike knew he neither had the time nor the patience to commit to that every three months, which led him to search for a better solution.

The previous rostering system at Whanganui Hospital was a patchwork of different tools:

The lack of a single source of truth meant constant confusion about start times, shift changes, and who was actually working when. Staff had to physically come to the office to check the wall planner before applying for leave.

Whanganui Hospital's radiology department manages:

Key challenges included:

After the initial roster was published, staff would immediately begin making swaps, essentially undoing the week of work that went into creating it.

When Mike analyzed the actual time spent on rostering, the numbers were staggering:

Building Rosters:

Managing Shift Swaps:

Total: 170+ hours annually - and that's not counting leave management, the constant checking, and the time individual staff spent managing their own schedules.

The cost was three times more than RosterLab's annual subscription, even including Mike's time to use the platform. It was presented as a pure cost-saving measure and was approved immediately.

The business case was straightforward: The annual time cost (170 hours × hourly rate) was three times the cost of RosterLab's subscription, including Mike's time to use it.

Hurdles faced in 2023:

What helped:

Mike notes: "By 2025, AI adoption has become much easier. Everyone uses ChatGPT now, people use AI trading bots, AI for music creation - AI is everywhere. I just happen to use it for rostering."

Initial Setup Time:

What the setup involved:

Mike's Advice:

"Put in as much time as you can possibly commit at the start. Get those rules right from the beginning. It'll save you tweaking time afterwards once you generate a roster."

The Learning Curve:

Initially, Mike spent time tweaking rosters after generation. However, as RosterLab's understanding of radiology rostering improved and they rolled it out to more departments, the process became significantly easier. Now, Mike barely has to do anything - the engineer handles most changes.

Roster Generation:

Shift Swap Management:

Total Annual Savings: 179+ hours

What Changed:

Unexpected Benefits:

Consolidated Systems:

All the separate systems (Excel, OneDrive, wall planner, paper forms, handwritten boards) were replaced with a single platform accessible from anywhere.

Multiple Views:

Communication Improvements:

Leave Management:

Equipment Scheduling:

Preference-Based Rostering:

Improved Transparency:

Shift Swapping Revolution:

Better Work-Life Balance:

Compliance & Fairness:

The philosophy shift: From "you get what you're given" to "balanced preference optimization while meeting service demands."

Mike's Top Advice:

"If you're going to use RosterLab, put in as much time as you can possibly commit at the start. Get those rules right from the beginning. It'll save you tweaking time afterwards once you generate a roster. That's the biggest thing I can say. And then just enjoy all the other benefits that you get from it."

What They Didn't Know They Were Missing:

Critical Success Factors:

Scalability:

From radiology, RosterLab has expanded to:

The Reality Check:

Mike is adamant: "If RosterLab disappeared, I'm not going back to Excel. I'm giving up rostering. I'm not doing it manually. No way."

Before Publishing:

Staff think about what they need to do in the future and input their preferences into the app. The system achieves 70-75% preference fulfillment with the current rule setup.

After Publishing:

The shift swapping is managed entirely through the app - staff can see who's working, propose swaps, and if both parties approve and no rules are violated, it updates automatically without manager involvement.

Mike's Experience at Whanganui Hospital:

Started with radiology, then expanded to:

The system works for different clinical teams as well as administrative teams. Each team can have completely different union rules, shift patterns, and requirements.

Key Point:

Union rules are described in natural language within the system. You can describe contract rules (e.g., "80 hours per two weeks"), fatigue rules, and make them shift-specific or team-specific. As long as you can describe it, the system can handle it.

Initial Setup:

At least 30 hours of dedicated time:

What Makes It Easier Now:

When Mike started, RosterLab was still learning radiology rostering. Now they've implemented it across multiple radiology departments globally, so the process is much faster. Mike's advice: "I probably should have spent more time initially, but our service went through major changes anyway, so we ended up rebuilding the rules."

Setup Complexity Varies:

Some rosters are naturally easier to set up than others. It depends on the specialty and level of complexity in your rules and requirements.

Mike's Recommendation:

"Put in as much time as you can possibly commit at the start. Get those rules right from the beginning. It'll save you tweaking time afterwards."

Yes - Open Shift Distribution:

When a shift gap is identified, managers can:

If Staff Accept Then Change Their Mind:

If someone accepts an open shift but later needs to reject it (due to life circumstances), they can reject it through the app. The shift automatically becomes available again for other eligible staff to pick up.

Notifications:

All eligible staff receive a notification ("Bing on their phone") when an open shift becomes available. They can see shift details and accept or decline directly through the app.

For Staff:

For Managers:

Coming Soon:

A leave planner in calendar view showing:

Current Status:

RosterLab doesn't currently have a direct integration with UKG, but it's being actively explored.

Workaround Available:

You can export roster data from RosterLab and import it into UKG. This ensures all roster data is ready for payroll processing.

Whanganui's Approach:

They use manual time sheets where each staff member writes their own timesheet, gets manager approval, and it's automatically sent to payroll. This is separate from the rostering system.

Note: The export/import method provides a bridge until direct integration is available, ensuring roster swaps and final schedules flow to payroll systems.

Mike's Experience:

Whanganui Hospital has two satellite radiology sites in rural areas, about two hours' drive away. They have fixed staff who work there, but when those staff are on leave, Whanganui staff travel to cover.

How They Handle It:

Other Use Cases:

RosterLab has been used for flying doctors who travel between multiple locations. The system can handle:

Pricing Structure:

Why Per-Staff Pricing?

As departments get bigger, the complexity increases significantly:

ROI Consideration:

At Whanganui, the cost of the subscription (including Mike's time to use it) was one-third of what they were spending on manual rostering time annually. The business case was straightforward and approved immediately.