Engineered parts are among the most complex items to procure. Their technical specifications are hidden in PDF drawings, knowledge often resides only in the minds of individual employees, and ERP systems provide little usable information for well-founded price negotiations. In this webinar, Tacto demonstrates how industrial AI closes this gap: from automated extraction of geometric features from technical drawings to data-driven target price calculation.

‍

In the webinar, Jan Scholich (Product Manager at Tacto) and Armand Gall (Implementation Manager at Tacto) explain how Tacto Intelligence automatically analyzes technical drawings, makes items comparable, and derives concrete target prices from them.

The Data Problem with Engineered Parts

‍

Engineered parts present procurement with a fundamental data problem. In the ERP system, only the item number, short description, and price are stored. The actual cost information, such as geometric complexity, material, surface treatment, or tolerances, is contained within the technical drawing itself. Five items labeled "corner elbow" can be purchased from different suppliers at completely different prices, without procurement being able to determine if the price differences are justified. The cost knowledge needed to classify such differences often exists only in the minds of experienced buyers. It is not mapped in the system and cannot be scaled. This is precisely where Tacto comes in, by creating a structured database for engineered parts.

‍

From Drawing to Digital Item Profile

‍

Tacto Intelligence automatically extracts relevant properties from technical drawings and converts them into a structured item profile. In a first step, referred to as "Level 0," basic properties are captured: material, weight, volume, surface treatment, holes, surfaces, and fits. Each item receives a vector that numerically maps its technical characteristics, making it comparable to other items. Furthermore, the drawing itself is stored as a visual embedding. This captures geometric details that go beyond the extracted individual features, such as tolerance classes and form complexity. This creates a comprehensive digital profile for each engineered part.

‍

Identify Similar Parts and Calculate Target Prices

‍

Based on the extracted item profiles, Tacto performs an analysis across the entire item base. Similar parts are identified, price differences are made visible, and put into relation with technical differences. The savings potential grows disproportionately with the number of analyzed parts, as more comparison opportunities arise. For target price calculation, Tacto Intelligence also incorporates context data from SRM: supplier performance, complaint rates, and current capacity information. External market data is also included, such as the development of raw material prices like the LME nickel price. From this, the system derives recommendations for action and proactively alerts procurement to potential for renegotiation. Important to note: Tacto does not see itself as a theoretical should-costing tool, but rather derives target prices from real transaction data and market developments.

‍

Bridge Between Engineering and Procurement

‍

Another aspect of the webinar is the cross-departmental use of drawing data. The analysis not only reveals savings potential for procurement but can also create added value in engineering. Automated common part recognition helps engineers identify existing parts instead of designing new ones. This reduces part variety and optimizes procurement costs from the outset.

‍

Conclusion

‍

This webinar demonstrates how industrial AI systematically unlocks previously hidden knowledge from technical drawings. By combining automated data extraction, cross-item comparisons, and external market data, well-founded target prices are generated, providing a new data foundation for purchasing engineered parts.

Armand Gall and Jan Scholich from Tacto explain how industrial AI automatically analyzes technical drawings, extracts geometric features, and derives target prices for procurement. The webinar illustrates the journey from data extraction and similar-part recognition to proactive recommendations, highlighting how engineering and procurement benefit from a shared data foundation.