Setting Up Your Workforce Team Mix Optimisation
Thu May 08 2025
Author
Alok Joshi
Optimisation with AI uses statistical models to find the best possible solution given a set of parameters to work in. All optimisation algorithms require data inputs, an objective to be reached, hard constraints to limit the search space of possible solutions and decision variables that can influence the outcome, but are in your control.
When it comes to optimising the team mix of your workforce (i.e. the full-time, part-time, casual and contractor cohorts), defining the right parameters can set you up to find the best solution from a possibility of hundreds or thousands of combinations of team mixes that could be selected.
Below are some considerations of how you might setup your objectives, decision variables and constraint parameters in this problem:
1. Define your objectives
What defines a successful team mix?
When considering the mix of team members that could make up your roster for the day / week / month, there are potentially many different options to optimise for. Here are some considerations when defining the objective of your team mix:
- Is it all about cost? Minimising cost that are associated with direct labour hire can certainly be key priority, but there can be other objectives to consider for example, you may be expecting an extended period of high volume trade and therefore maximising efficiency at the cost of hiring more people may be acceptable.
- Do you need to be flexible with your objectives? Think about how often you think you would be changing the measures of success. Do you need to define different objectives at different times of the year depending on trade patterns?
This can be a balancing act, but needs to be defined upfront to setup any recommendations and simulations up for success.
2. Set Your Decision Variables
What am I allowed to change?
Decision variables are the parameters that you can control and adjust. They are within your remit to change and what the algorithm needs to consider to optimise with all the data it is provided.
Some considerations of decision variables involved in workforce planning include:
- Honouring the shift preferences of your employees: think about shift lengths, days of the week, weekend work, overtime.
- Catering for department specific needs: are you aware of any special training that's needed? Can you adjust for production times? Are there unique peak trading hours?
- Measuring your performance in other ways: what are the other metrics you use to determine if you have the right mix? Do you assess impact to NPS / customer satisfaction? Service level agreements? Safety impacts?
These can all be set as part of acceptable thresholds for your algorithms, defined by you, that can ultimately impact how well you meet your overall objective set.
3. Set Your Constraints
What are my limits?
Defining the non-negotiable rules will ensure the algorithm only recommends what is feasible and legal, especially when it comes to honouring employee agreements.
Examples of constraints that must be considered include:
- Compliance to any changes to legislation or bargaining agreements
- Operational and business rules that restrict certain activities
- Employee restrictions in terms of skills/abilities
Once these are setup, it is then up to “the maths” i.e. the algorithms to determine the best combination to meet your objectives, respecting the constraints as the limits and decision variables it must operate within. This is where the power of simulation can also be very useful i.e. change variables like acceptable ratios and operating hours, see the impact to your objective, compare team mix outputs, save out scenarios for responsive planning in times of disruptions etc.
Take it one step further with an AI Twin and treat your forecasted sales as in input, predicted inventory levels, generated order plans, and you could truly make decisions that consider end-to-end impacts.
If you need help in establishing your team mix optimisation, reach out to one of our experts (info@jahan.ai) to discuss what defines a successful team mix for you.
aiml
aiTwin
retail
roster
workforceOptimisation