AI assistant Significantly accelerate construction times

A guest contribution by Pedram Shahid | Translated by AI 4 min Reading Time

Related Vendors

Ever shorter construction times and at the same time increasing demands for innovative products—that sounds like two goals that cannot be reconciled. But an AI assistant can solve the dilemma.

Through a chat box in the CAD software, designers address the AI assistant.(Image: CADAICO)
Through a chat box in the CAD software, designers address the AI assistant.
(Image: CADAICO)

Pedram Shahid is the managing director of CADAICO GmbH

Cost pressure and innovation are also omnipresent in mechanical engineering. However, when used correctly, artificial intelligence (AI) has great potential to make the processes in product development more efficient and innovative. One solution for this is the Cadai Assistant.

The AI assistant for design has a decisive advantage: it is specifically trained based on the company's CAD data—rather than on generic design data—while also ensuring the protection of sensitive information. This allows the design time of even complex assemblies to be reduced by 40 to 60 percent, leaving users valuable time for truly innovative design tasks. The assistant currently functions as an add-on for Solidworks and is planned to be integrated with other relevant CAD programs in mechanical engineering in the future.

High effort leads to high costs

We all know the problem: engineering efforts significantly affect product costs. The engineering efforts are largely caused by manual activities. The resulting engineering costs are directly reflected in the product costs. Especially for customized products such as machines, machine components, and equipment, engineering efforts can account for up to 70 percent of product costs. In particular, in countries with high labor costs such as Germany, Austria, or Switzerland, this leads to a significant cost disadvantage compared to competitors from low-wage countries.

The more manual activities involved in a product, the greater the cost disadvantage for these companies. Every year, mechanical engineering companies lose valuable market share because they are no longer price-competitive.

Perfect conditions for AI

Despite all improvements and standardizations, designers still spend a lot of time developing variants of existing products or assemblies with a high manual effort. Sometimes assemblies are completely newly designed because they differ in so many aspects from existing solutions that a redesign is not worthwhile. Yet, in most companies, there are enough CAD data present, and the requirements for new variants are often sufficiently well described, allowing large parts of the design to be automated. Thanks to the rapid development of current AI technologies, we are now able to implement exactly this.

The previous AI solutions are trained on general data and have nothing to do with a specific company's products. Yet, especially in mechanical engineering, in-house product knowledge is essential—companies often invest decades in developing unique features and innovations to set themselves apart from the competition. A provider of hydraulic cylinders, for example, does not want to design arbitrary hydraulic cylinders, but rather exactly the hydraulic cylinders with the product features of their company.

This is how the Cadai Assistant works in CAD

This is where the Cadai Assistant solution comes into play. For each company, a separate instance of the AI assistant is set up, which is then trained exclusively with the company's own CAD data, thus perfectly representing the products and product features. This results in a very important security aspect: the design and product knowledge remains in the company's own instance of the assistant and thus within the company. Moreover, there are no systemic or procedural minimum requirements for implementing the solution. Currently, users of the CAD software Solidworks benefit from an interface in the form of an add-on, but further interfaces to common CAD programs are planned. This way, designers work in their familiar CAD solution and launch the assistant directly in their CAD software.

  • The Cadai Assistant opens as a chat window within the CAD program and can now perform design tasks.

  • The use is very intuitive: users type into the chat window what design task they need, for example: "Create a hydraulic cylinder with dimensions XYZ, for a load capacity of X kN based on the cylinder type ABC100. Use the calculation formulas in our design guidelines for the layout."

  • And the assistant immediately begins the design. Within seconds, it creates a fully functional 3D model in CAD based on the order and the company's own data.

  • This model is created in the native file format and contains all the properties so that it can be further processed as usual.

  • Language, correct grammar, or spelling are not relevant: The assistant supports almost all languages and is fault-tolerant.

    Subscribe to the newsletter now

    Don't Miss out on Our Best Content

    By clicking on „Subscribe to Newsletter“ I agree to the processing and use of my data according to the consent form (please expand for details) and accept the Terms of Use. For more information, please see our Privacy Policy. The consent declaration relates, among other things, to the sending of editorial newsletters by email and to data matching for marketing purposes with selected advertising partners (e.g., LinkedIn, Google, Meta)

    Unfold for details of your consent

No vast amounts of training data needed

It is often assumed that huge amounts of data are needed for training. For earlier AI methods, this concern was indeed true—enormous amounts of data and months of training were necessary before the algorithms were operational. This is where the Cadai Assistant differs: the solution can be trained with just a single CAD file of an assembly or product. Thus, the company's own assistant is ready within a very short time to generate different variants of this assembly.

Moreover, a major advantage is that the solution designs complex assemblies or machine components—and not simple parts that designers can model themselves with a few clicks and without much time effort. A user was able to reduce the design time for an assembly from one week (new design) to just one day. This results in an enormous productivity potential of 40 to 60 percent. A potential that can be better utilized to be more competitive and future-proof.