bg
Digital products and platforms
19:08, 16 November 2026
views
1

Russia Develops AI Agent to Optimize Engineering Drawings

Russian developer and systems integrator of engineering software CSoft Group has introduced an AI agent that can automatically improve drawings based on 3D models. The model and its training algorithms were developed in-house, without relying on third-party technologies.

One notable advantage of the digital solution is that it can run on a design engineer’s workstation, keeping the virtual assistant from transmitting engineering data to external services. Its hardware requirements do not exceed those of a workstation running Model Studio CS, and it does not require a dedicated server. This is particularly important for companies and industrial enterprises that work with engineering documentation.

Improving Drawing Readability

According to CSoft Group, optimizing drawing callouts is one of the tasks clients request most often. The reason is straightforward. When a design engineer automatically generates drawings from an information model, document preparation takes less time. But automated placement can create a problem: on sheets densely packed with equipment, piping and structural elements, callouts may overlap one another or obscure the drawing graphics. Design engineers typically have to reposition them manually to make the drawing legible and ensure it meets drafting standards. This is not a simple task. Moving even one annotation can make the placement of several neighboring callouts worse. The design engineer must account for all these interdependent constraints and find a workable arrangement for the entire sheet.

To optimize callout placement, the developer's team created an AI agent based on reinforcement learning. The algorithm analyzes the positions of annotations and moves them to produce a more readable drawing. It operates directly in the drawing environment and then receives a score based on specified criteria: no overlapping text, no intersecting leader lines, no crowded annotations and no elements extending beyond the sheet boundaries. CSoft Group's experts developed both the model and its training algorithms in-house.

A First-of-Its-Kind Solution for Engineers

According to the developer, this is the first AI-powered solution in Russia designed to automatically optimize the placement of annotations on engineering drawings. CSoft's measurements show that the agent can reduce the time needed to complete drawing layouts by up to 2.5 times on complex sheets. The actual savings depend on how densely the drawing is laid out and how many conflicts it contains. In the first version of a drawing, the system resolves most cases in which text callouts overlap one another or collide with other drawing elements. In subsequent development stages, it will eliminate intersections between leader lines and then address other types of annotations.

This step-by-step approach allows the agent's capabilities to expand incrementally, rather than limiting it to a single drafting operation. Roman Vyatkin, head of CSoft Development's Ivanovo branch, explained that the team treats callout placement as a sequence of decisions. The agent first receives a drawing, then moves annotations within the document and learns from the results of its actions.

A similar approach has previously been used to place labels on maps, but no ready-made application was available for engineering drawings. CSoft's developers filled that gap by designing their own model architecture and training it to account for the specific requirements of engineering documentation.

Integrating AI Into the Design Workflow

For the company's clients, control over engineering data and predictable tool performance are especially important. That is why CSoft's specialists are developing their own AI tools, which run within the user's environment and fit into standard design workflows.

CSoft's AI agent is not a general-purpose chatbot. It addresses a specific engineering task and has become part of a professional design workflow. If the solution proves successful in broader use, demand for applied AI in computer-aided design (CAD) and building information modeling (BIM) could grow in Russia. Over time, software designed to improve drawing readability could help engineering departments at industrial companies, architecture and construction firms, and engineering consultancies while reducing the risk of errors. Wider industrial adoption could also save substantial time and labor. Even a single project can generate extensive documentation that must go through lengthy review and approval processes and collect multiple signatures.

Artificial intelligence technologies are widely used in engineering, particularly in areas such as documentation and requirements management, expert systems, computer vision and generative design. These are practical applications that are in high demand across a range of industrial sectors
quote
like
heart
fun
wow
sad
angry
Latest news
Important
Recommended
previous
next