Master variant management: AI-supported optimisation with the FIRE Engine

A diverse product range creates opportunities for increasing profit margins and using company resources more effectively. However, it also increases complexity within the company. This means that the associated costs increase disproportionately in relation to the company's total costs.

At the same time, strong competition is forcing companies to both significantly reduce their costs and increasingly offer customer-oriented, i.e. more individual and higher-quality products.

In this context, it makes sense to take a closer look at the topic of variant management in companies. This helps you to reduce and control existing complexity and to avoid potential future complexity in your company.

To this end, I would like to present the AI-supported similarity analysis, which makes variant management more efficient and can significantly reduce the associated costs. It identifies similarities between product variants or components and suggests a more cost-effective variant to users. With the help of the FIRE Engine, which supports this modern method, companies can efficiently manage their variant diversity and optimise the configuration process for their customers.

I will also be explaining the exciting possibilities that technologies such as artificial intelligence (AI) and large language models (LLMs) open up within the FIRE Engine.

What is the ‘FIRE Engine’?

The FIRE Engine is a framework for efficiently modelling simple, complex or multi-level configuration models, with which all conceivable configuration scenarios can be implemented very quickly and easily – from feature-based top-down product configuration to complex bottom-up plant design or solution configuration.

What is meant by ‘similarity analysis’?

Similarity analysis is a method for examining product diversity. It involves comparing individual elements, such as product variants or components, on the basis of their characteristic features.

To do this, the user can define a tolerated ‘equality’ with regard to the characterising features between the desired component and a standard component. For example, two components are ‘equal’ if they differ by no more than two centimetres in the ‘length’ property. This comparison is not limited to components within a product, but can also be applied to products that have already been configured. This means that standard components can be reused, redundant variants avoided and the diversity of variants consolidated in a targeted manner. This enables the manufacturer to use resources more efficiently and helps the customer to opt for a lower-priced product variant that also meets their requirements. This significantly increases both the customer experience and customer satisfaction.

How does the FIRE Engine support similarity analysis?

The FIRE Engine uses modern algorithms that enable a precise analysis of product data. It supports exactly this ‘recognition of similarities’ in order to use them specifically for optimisations.

A concrete example of how this can be used is to improve the Engineer-to-Order (ETO) approach by using standard components. In the ETO process, in which individual product solutions are developed, the similarity analysis of the FIRE Engine supports the identification of standard components that can be adapted to the specific requirements of the customer or already meet them. This leads to a price-efficient component selection while simultaneously reducing complexity. This creates a solid basis for mapping customer requirements more precisely and satisfactorily.

AI-supported configuration processes and the potential of LLMs

To make similarity analysis even more efficient, we also rely on AI and LLMs. FIRE Engine can realise its full potential when used in combination with these technologies.

• 1. Automated component/assembly optimisation: With our expertise, we can show you how similarity analysis can be made even more efficient with AI. For example, we can automatically incorporate empirical values from production or logistics, analyse them and suggest similar components or configurations to the user.
• 2. Voice-controlled variant configuration: In addition, we help you integrate the potential of LLMs into your processes. Together, we can develop solutions that make it possible to ask questions in natural language such as: ‘What similar parts are available for this configuration?’, ‘What workflows need to be triggered to produce this part?’ or ‘Find the most cost-effective variant for this component!’.

We not only offer you a modern CPQ system, but also automate your entire end-to-end process using AI. Our systems automatically search your PLM or ERP databases, identify parts that have already been manufactured and suggest the most efficient workflow. With our support, we turn this vision into reality and create a system that sustainably optimises your processes.

Why is AI-supported similarity analysis indispensable?

The increasing variety of variants in the manufacturing industry makes the use of tools to reduce complexity while simultaneously increasing efficiency indispensable. Similarity analysis in conjunction with AI enables targeted control and optimisation of product diversity and is therefore a central building block for meeting the challenges of the modern manufacturing industry and for positioning yourself strongly and sustainably in the face of competition. Companies that rely on the FIRE Engine benefit from a leaner product structure and reduced production costs. In combination with other FIRE Engine functions, such as instantiation of components, plausibility checks and comprehensive support for the ETO process, the framework offers a comprehensive solution for variant management.

Thanks to the robust framework and future-oriented technologies such as AI and LLMs, the FIRE Engine not only offers your company a solution to current challenges, but is also a valuable investment in the future.