It's Thursday afternoon and you've half-built next week's roster. You're staring at a spreadsheet, trying to work out how busy Saturday night will really be.
Book too many people and your wage bill increases. Book too few and your team is under pressure while customers wait.
That balancing act is exactly why more venues now use AI labour forecasting for rosters. Deputy's forecasting and Smart Scheduling tools can use business data to help estimate demand and staffing requirements. Deputy AI (artificial intelligence) can separately assist managers with supported scheduling tasks. That way, you plan cover before you publish, rather than reacting once a shift runs into problems.
You've also got real constraints in the mix, like award rules, penalty rates, and last-minute shift swaps, which make guesswork expensive. Adoption is climbing fast: 78% of organisations report using AI in some form, according to Stanford's Human-Centered AI Institute.
Australian shift workers are keen too. 54% of Australian shift workers say they want more AI support at work, compared with 46% in the United States. That figure comes from Deputy's 2026 Australian Big Shift Report, which drew on more than 60 million shifts worked across Australia.
This guide walks through what AI can and can't do for your rosters, and the data it needs to work. It also covers how to read a healthy labour cost percentage, award considerations in Australia, and a simple way to get started.
The theme throughout is straightforward. AI can assist with scheduling tasks, while forecasting and Smart Scheduling tools provide data-driven insights. Your managers review the information and make the final roster decisions.
Key takeaways
Deputy Forecast and Smart Scheduling can support more data-informed roster decisions, reviewed by a manager before you publish
A forecast is only as good as the data behind it, so clean, connected data matters more than any single feature
Smart Scheduling helps managers understand forecast demand, scheduled wages, and labour costs before you publish a roster
People still own the final call, fairness, and compliance; AI is the assistant, not the decision maker
How AI labour forecasting for rosters improves your staffing decisions
There's a lot of noise around AI, so it helps to be clear about the real, day-to-day wins. The point isn't to replace your judgement. It's to hand you a stronger starting point, so you spend less time wrestling with numbers and more time with your team and customers.
Australian venues are feeling cost pressure from rents, insurance, utilities, and wages. The Australian Bureau of Statistics (ABS) Wage Price Index rose 3.2% over the year to the June 2026 quarter. Many venues are shifting from a hire-to-grow mindset toward optimising the team they already have.
Deputy's 2026 Australian Big Shift Report describes this move toward leaner, tighter rosters. Better forecasting can help protect margins without cutting service.
Build rosters around expected demand, not guesswork
Instead of reviewing last week's numbers, Deputy's Smart Scheduling tools can use demand metrics such as sales, transactions, and foot traffic to inform staffing decisions. Managers can also apply their own knowledge of factors such as seasonality, weather, and local events when reviewing the roster.
You still know things the data doesn't. A big local game, a supplier delay, or a staff member's exam week all shape the roster. The forecast gives you a sensible draft, and you adjust it with what you know on the ground.
Project labour costs before you publish the roster
One of the most useful moves is costing a roster before you roster anyone on. As you build shifts, Deputy's scheduling insights can help managers compare scheduled wages against expected sales before publishing a roster. That way you can spot an over-budget Saturday while there's still time to change it.
That wage-versus-revenue view helps managers make more informed decisions about where to add or trim hours. It won't make the call for you, but it takes a lot of the blind spots out of planning.
What AI labour forecasting can't do on its own
AI isn't a silver bullet. A forecast is only as reliable as the information behind it, and it can't read the human side of a shift. Knowing where the limits sit is what keeps your rosters trustworthy.
Why your data quality makes or breaks the forecast
If your data is scattered across point-of-sale systems, spreadsheets, and disconnected apps, the forecast inherits every gap and error. Inconsistent inputs lead to shaky suggestions, and your team notices quickly when the roster feels off.
Clean, connected data is the foundation. When sales history, past rosters, and availability all live in one place, the forecast has a fair chance of being useful. Garbage in still means garbage out.
Why managers still own the final roster call
AI can surface a pattern. But it can't tell that one team member is new and needs a steadier shift, or that two people work better apart. That context lives with your managers, and it's what turns a draft into a fair, workable roster.
Deputy's 2026 Australian Big Shift Report frames the future as AI-assisted, human-responsible rostering, where AI supports the decision and people stay responsible for the outcome. As the report puts it, technology can improve how people perform work, but management shapes how work feels.
AI forecasting versus rules-based and manual rostering
Most shift-based businesses use one of three approaches. Each has a place, and it's worth knowing where AI-assisted forecasting actually earns its keep. Keep in mind that forecasting can give managers a data-driven input alongside their own operational judgement.
The honest summary: manual works until it doesn't, rules-based handles the steady stuff, and forecasting and Smart Scheduling help most when your demand moves around. Forecasting and labour modelling use configured business data and metrics to inform staffing requirements, while Deputy AI assists with supported scheduling tasks. In all three, you're still the one who signs off the roster.
The data your labour forecast and AI-assisted roster need
A forecast is a reflection of what you feed it. The more complete and connected your inputs, the more useful the suggestions. Most venues already have this data; it's just sitting in different corners.
Forecasting and labour modelling: historical sales, sales forecasts, transactions, foot traffic, and other configured demand metrics.
AI-assisted scheduling: relevant information already available within Deputy, plus the scheduling request you provide as the manager.
You don't need all of this on day one. Start with sales history and past rosters, then add availability and events as you go. Each input you connect gives the forecast a clearer picture to work from.
People plus AI: the better together approach
Forecasting works best when AI and managers work in tandem. AI brings speed and pattern-spotting across mountains of data.
Your managers bring empathy, context, and the read on their team that no tool can replicate. Put them together and AI labour forecasting for rosters can move from reactive to more planned.
How AI-assisted tools can support roster decisions
When managers need a faster scheduling starting point, tools like Deputy AI can generate proposed changes for review. Deputy AI can help authorised managers build schedules, edit shifts, fill open shifts, and publish schedules using natural language. Deputy's forecasting and Smart Scheduling capabilities handle demand forecasting and labour modelling. It doesn't replace the human call; it can give your managers a faster, better-informed starting point.
In practice, that can mean smarter rostering suggestions based on real business data. You also get a clearer view of upcoming demand against labour cost.
Configurable rules and alerts can help you spot potential issues early. Managers stay in control, adjusting as they go.
For an operator's view of what that can free up, here's General Manager Mari Bornelli from a Sydney hospitality venue:
Each pay cycle with the previous system, I was spending around 2 to 3 hours to do payroll, and now with Deputy it takes me 45 minutes. My level of compliance confidence was pretty low at about 50%, and I'm at 80 to 90% now.
Mari, General Manager, Funk Drinks Co.
Results like these depend on your setup, so treat them as an example rather than a promise. The pattern, though, may be less admin and more confidence in the roster, depending on your setup.
