LAD AI Assistant Brings Conversational Interface to SolidWorks CAD Workflows

SolidWorks users spend significant time hunting through toolbars and remembering exact parameter names for routine operations. A startup called TryLAD is addressing this friction with an AI assistant that translates plain English commands into SolidWorks actions.

The core concept is straightforward: type "create a 50mm fillet on all edges" or "extrude this sketch 20mm in both directions," and LAD executes the operation directly within the SolidWorks interface. This bypasses the traditional workflow of selecting specific tools, adjusting property managers, and manually entering values through multiple dialog boxes.

What makes LAD different from generic AI coding assistants is its deep understanding of CAD-specific terminology and SolidWorks' feature tree structure. The assistant recognizes that "fillet" means a specific geometric operation with radius parameters, not just a generic 3D modeling concept. It can interpret ambiguous requests like "make this part symmetrical" by analyzing the current selection context and proposing appropriate mirror or pattern operations.

The technical implementation appears to use a combination of natural language processing and SolidWorks API integration. Rather than attempting to replace the native interface, LAD operates as an overlay that monitors user input, parses commands, and programmatically triggers the equivalent menu selections and parameter entries. This approach means the AI doesn't need direct access to modify SolidWorks' core code—it simply automates the interface interactions that a human would perform.

For complex assemblies, LAD can chain operations together. A request like "create a housing for these components with 3mm walls and mounting holes" would involve multiple steps: analyzing component dimensions, generating a bounding volume, creating the shell, then adding hole features based on mounting point locations. The assistant maintains context about the active document, selected entities, and recent operations to make intelligent decisions about what "these components" refers to.

The tool also handles parameter relationships. If you tell it to "make the slot width match the shaft diameter," it can establish a dimensional constraint that updates automatically when the shaft changes. This goes beyond simple command execution into actual design intent preservation.

Early adopters report the biggest value in repetitive tasks. Setting up standard part templates, creating feature patterns, or applying consistent fillet treatments across multiple components become single commands rather than multi-step processes. The time savings compound when working with large assemblies where finding the right feature in the tree can take longer than the actual modeling operation.

However, the assistant isn't a replacement for CAD expertise. It can execute commands but won't catch design logic errors or suggest better modeling approaches. A user still needs to understand what they want to achieve and why. The AI helps with the "how" but not the "what."

LAD joins a growing category of AI tools targeting specific professional software ecosystems. While general-purpose AI assistants can help with code or writing, domain-specific tools like LAD aim for deeper integration and more precise command execution. The approach mirrors how GitHub Copilot specialized in code completion rather than trying to be a general writing assistant.

The business model follows the pattern of other professional tools: subscription pricing based on usage tiers. For engineering firms where CAD time directly translates to billable hours, even modest efficiency gains can justify the cost. The question is whether the assistant's accuracy and reliability match the precision requirements of engineering work, where a misinterpreted command could introduce costly errors.

TryLAD is currently available as a beta for SolidWorks 2022 and later versions. The company is collecting feedback on command interpretation accuracy and expanding its library of supported operations beyond the core modeling features currently covered.

For teams considering adoption, the key evaluation criteria should be command success rate on your specific workflows, integration stability with existing SolidWorks customizations, and the learning curve for team members accustomed to traditional interface navigation. The tool promises speed, but only if it reliably understands what you're asking for.