From Technical Drawings
to Intelligent Part Recommendations for Facil

For Facil, identifying the right fastener based on a customer’s technical drawing was a manual and knowledge-intensive process. Engineers had to interpret complex engineering drawings, extract technical specifications and compare them against thousands of catalogue parts to identify suitable alternatives.

We’ve developed an AI-powered solution that automates this entire workflow, enabling faster proposals, greater consistency and improved scalability.

AUTHOR – Josse

The challenge

Matching a customer’s technical requirements to the right catalogue part is far more complex than a simple product search. Engineers and sales specialists at Facil spent significant time manually interpreting technical drawings, extracting specifications such as length, diameter, thread pitch, head style, coating and material, before searching their catalogue for suitable alternatives.

With hundreds of drawings arriving every week, this manual approach became a bottleneck. It was time-consuming, prone to inconsistencies and difficult to scale. On top of that, the expertise required to correctly interpret drawings and recommend alternative parts largely resided with experienced employees, making it difficult to transfer and preserve this valuable knowledge.

Our solution

We’ve developed an end-to-end AI solution that automates the entire journey from technical drawing interpretation to alternative part recommendation.

Using a vision-capable AI model, engineering drawings are automatically analysed and key technical specifications are extracted within seconds. These attributes are then matched against Facil’s product catalogue to generate a ranked shortlist of suitable alternatives, including a clear explanation whenever deviations from the original specification occur.

Our solution continuously improves over time through user feedback. Validated recommendations help refine both the extraction accuracy and the ranking quality, while an integrated testing framework ensures every improvement is objectively measured against a ground-truth dataset before deployment.

“From manual drawing interpretation to intelligent part recommendations.”

The impact

Our solution frees Facil’s engineers and sales specialists from the most repetitive and time-consuming parts of their workflow, allowing them to focus on understanding customer needs and delivering value.

The solution delivers:

• Faster proposal turnaround by processing multiple technical drawings simultaneously and returning ranked alternative parts within seconds.

• Greater consistency through a structured, AI-driven extraction and matching process.

• Increased scalability without the need to proportionally increase headcount.

• Preservation of valuable engineering expertise by embedding domain knowledge into a reusable and auditable AI solution.

This project demonstrates how vision AI, combined with strong feature engineering and continuous learning, can successfully automate highly specialised industrial processes. Rather than replacing engineering expertise, we’ve created a solution that amplifies it, enabling Facil to operate faster, smarter and more competitively.