Advancing digitalisation is having a major impact on the food industry, especially confectionery production. OMMM has developed an innovative AI application for production planning to meet the increasing demands and boost efficiency. Artificial intelligence (AI), especially expert systems, is seen as the way forward and has already shown positive results in the food industry. OMMM is an up-and-coming startup specialising in standardised AI-based planning software. OMMM's ultimate goal is to provide customers with a measurable return on investment (ROI) by optimising supply chain planning with AI.

AI, defined as the ability of machines to perform tasks that normally require human intelligence, offers a variety of benefits in production planning. AI expert systems, such as those used by OMMM, use rules and heuristic methods to solve problems and complement human expertise. In the food industry, AI can be used in various areas of the supply chain to increase efficiency, improve product quality and reduce costs.

The challenges in production planning, especially in the food industry, are diverse and complex: Customer and internal orders must be fulfilled, taking into account numerous constraints and restrictions. These include technical and personnel resources, minimum and maximum batch sizes, sequence-dependent cleaning processes, allergen-related restrictions and best-before dates. In addition, planners are often faced with conflicting goals such as cost reduction, meeting delivery deadlines and capacity utilisation.

With the help of AI in production planning, planners are given the opportunity to objectively select the best plan from hundreds of thousands of possible plans in this complex environment. This results in massive cost savings for the company.

One of Europe's leading fruit processors has been using an AI-based production planning tool (PP module) for several years, which forms the basis for the current OMMM module. This multinational manufacturer employs well over 1,800 people in more than 40 countries. The initial situation: The manufacturer operates batch production on up to 35 production lines. In addition to short-term customer orders, there are also longer-term stock orders. Production draws on up to 2,000 raw materials, and the planning team initially consisted of 12 people who mainly worked with Excel spreadsheets. The planners' daily work was characterised by a highly complex planning situation, with daily rescheduling due to new customer orders. Generating an acceptable plan was almost impossible, let alone comparing several versions. This could only be managed by highly experienced and resilient planners. The plan generation took hours, the results were suboptimal, and the company was heavily dependent on the expert knowledge of individual planners. In addition, the efficiency of the planning (costs, OEE, material, cleaning) could not be measured.

The challenge was to ensure digitalisation, automation and integration into the existing IT landscape, to create a “frozen horizon” to stabilise planning and production, to achieve cost-optimised planning and cost transparency and at the same time to enable the possibility of short-term replanning and manual intervention.

The project was implemented in four steps: First, all relevant production and planning processes were analysed and documented. The relevant master and transaction data was then cleansed. Then the planning tool was trained to meet the specific requirements of the individual planning situation – accompanied by several test loops. Finally, the roll-out took place.

The results of the PP module – admissible plans optimised according to the individual specifications – are displayed on several large monitors in the planning office. These can be created at the touch of a button. These plans can be evaluated and are based on real-time data from the shop floor. The planner checks any warning messages that occur and releases the plan. The production orders are then seamlessly fed back into the leading SAP system. Planners can intervene manually at any time and evaluate their actions based on specific criteria.

In addition, the tool offers a visual representation of the results in the GANTT diagram as well as an evaluation and display of the planning results (production orders optimised in terms of date and quantity). The intuitive user interface is easy to use, and the quality of the plan can be evaluated at a glance using colour coding.

Planning is now possible during a night run, but also at any time during the day. The way the planners work has changed drastically. Bottlenecks are communicated proactively and the planning team can focus more on other topics such as OEE optimisation. The “frozen horizon” that has been introduced has calmed planning and production.

Overall, savings of around 1.2 million euros were achieved in the first year following the introduction of the project. The entire project has already more than paid for itself in the first year after its introduction – the potential cost savings in production planning are significant. The PP module is now being rolled out worldwide.

The head of production planning initially doubted whether an expert system could adequately handle the complexity. However, he is now enthusiastic about the OMMM software. Matthias Banike, VP Business Development at OMMM GmbH, emphasises: “Our goal is to achieve an ROI of less than 12 months for our customers and to turn the planners into enthusiastic AI optimisers.”