Eggsactly planned by AI. So that Switzerland doesn’t run out of eggs.
What does a transparent, reproducible and reliable AI solution for real-world breeding planning look like? This is precisely what BWO Systems AG demonstrates with its Breeding Planner Agent within the BWO Concierge platform.
At Swiss poultry specialist Prodavi SA, a specialised ‘Breeder Agent’ automates the complex breeding planning process. A task that, in many businesses, is still organised in Excel today with a great deal of manual effort.
When planning becomes a bottleneck
Breeding planning is not a theoretical optimisation problem, but a daily operational reality. Limited housing capacity, sanitary fallow periods, specific animal welfare label requirements, delivery deadlines and complex weekly schedules constantly influence one another. Even minor changes can trigger extensive adjustments across the entire production planning process.
Particularly in Swiss egg production, where reliability, animal welfare and transparency are top priorities, planning quickly becomes a business-critical factor.
In such scenarios, many companies still rely on Excel spreadsheets, empirical knowledge and manual correction loops. This works as long as complexity and time pressure remain manageable. In real production environments, however, this approach quickly becomes a bottleneck.
Prodavi SA, based in Schötz (LU), is a leading, family-run Swiss company specialising in the production and distribution of laying chicks and pullets.
The company covers the entire production chain, from parent stock and the hatchery through to the rearing and husbandry of pullets and laying hens. This role is central to the Swiss poultry industry because availability, quality and predictability are closely interlinked throughout the entire value chain.
At the same time, the pressure for stable processes is constantly growing: demand for Swiss eggs remains high, whilst regulatory requirements, quality standards and production safety are continually increasing.
The response from BWO
To address this very challenge, BWO Systems AG developed a specialised Breeder Agent based on BWO Concierge.
The solution automates breeding planning – from order allocation to validated weekly allocation – whilst taking all relevant restrictions into account simultaneously:
- Barn capacity
- Cycle duration
- Downtime periods
- Label requirements / certifications
- Parent stock management
- Allocation and stocking rules
The key point, however, lies in the technical architecture: the large language model is used exclusively to interpret freely formulated rules. The actual planning takes place in a controlled, rule-based manner within BWO Concierge – fully traceable, reproducible and deliberately not designed as a black-box AI.
That is precisely the key message of this project: industrial AI can be powerful without losing control.
An overview of the planning process
Before starting, you can choose between two modes:
- Option A – Optimise planning: An existing plan is tailored to specific requirements. Changes to customer or grower data, as well as parameters, can be incorporated flexibly and dynamically.
- Option B – Re-planning: The entire plan is completely recalculated and rebuilt based on the latest data.
The user process follows five clearly structured steps:
1. Import data
Breeder, barn, certification and order data are imported directly from existing Excel files. In ‘Plan-Only’ mode, the solution can even track and reuse previous decisions such as ‘OK’, ‘Not OK’ or ‘Ignore’.
2. Interpreting rules
A Large Language Model (LLM) interprets freely formulated rules from the parameter tables and translates them into specific planning logic – such as exclusions, priorities or automatic adjustments.
3. Calculate the project
Based on the delivery weeks, the system automatically calculates the overall rearing plan. In doing so, it takes into account rearing duration, barn downtime and available barn capacity, and assigns orders to the appropriate rearing units and barns.
4. Check for conflicts
Every schedule is automatically checked for weekly capacity, mandatory breaks and allocation rules. If a job cannot be scheduled appropriately, the system clearly flags it as ‘Not Plannable’, including the reason.
5. Provide results
Finally, the solution generates a multi-sheet Excel workbook containing planning data, capacity utilisation, orders and integrated master data. Business users can review and adjust the results and reuse them for the next planning cycle.
Measurable benefits
The greatest impact does not come from a single feature, but from the end-to-end planning process.
What used to take days or even weeks to plan manually in Excel can now be fully calculated and validated in a matter of minutes. According to the project data, this reduces the planning workload significantly
In addition, there is a significant economic benefit:
- better capacity utilisation
- more stable weekly planning
- significantly fewer manual correction cycles
- greater transparency in the event of conflicts
- reproducible decision-making criteria
According to project data, the documented return on investment is around 15 times the initial investment per year. The quality of the automated planning is also impressive: in the scenarios described, the solution typically achieves a success rate of 97 to 99%. Cases that cannot be planned for are not concealed, but are reported transparently, including the cause.
René Steiger, who is responsible for the breeding operations at Prodavi, describes the solution as a practical tool for long-term optimisation and planning several years in advance.
“The Breeding Planner makes our complex breeding planning manageable and easy to oversee. It gives us greater speed, transparency and significantly more certainty in our long-term production planning.”
René Steiger, Head of Breeding Operations at Prodavi
AI as a Service from Switzerland
The Breeder Planner Agent was developed by BWO Systems AG in Schenkon and is provided as a service via the BWO Concierge. This means that customers do not need to set up their own costly AI infrastructure, yet still have access to a productive, business-critical solution.
The Swiss location is a decisive advantage, particularly when it comes to sensitive production processes:
- Hosting in a Swiss data centre
- Data protection and digital sovereignty
- Controllable operations
- High availability and governance
This is how industrial AI works in practice: transparent, manageable and productive.
Experience BWO Concierge for yourself
BWO Systems AG doesn’t just talk about AI; it is already successfully deploying it in live customer solutions. The project with Prodavi SA serves as a prime example of how real-world industrial problems can be solved using intelligent yet controllable AI systems.
Anyone wishing to experience BWO Concierge and its wide range of specialised agents in action will have the opportunity to do so at AI WEEK 2026. The event will take place on 19 and 20 May 2026 in Milan (Fiera Rho) and is one of Europe’s leading events for practical AI applications.
Visit us in Milan or contact us directly to experience our AI solutions first-hand and discover the potential for your own processes.
.png)
