AI Advisor

Helping Engineers Understand Unfamiliar Code with Confidence

Validation engineers often inherit VIs without documentation, making it difficult to understand what the code does or how to reuse it. This case study illustrates how the AI Advisor supports engineers by generating clear, contextual explanations, turning a frustrating task into a faster, more confident workflow.

Situation

Test and measurement customers use our software to run automated measurements, often relying on shared or legacy VI’s (Virtual Instruments) written by others. This code is rarely well-documented, making it hard for validation engineers to quickly understand what it does or how to adapt it. This slows down workflows and increases the risk of errors. We saw an opportunity to use AI to help engineers interpret and trust unfamiliar measurement code and repurpose it for their bespoke needs.

Task

These engineers need to quickly understand unfamiliar code, often written by someone else, so they can repurpose it for their own testing needs. Their task is to interpret what the code does, how it works, and whether it fits their specific context, all without spending excessive time or relying heavily on others. Most importantly, they need to do this without being experts in software development.

Action

When the user opens a VI in LabVIEW, they can activate our AI advisor, to help interpret the code. There are a few mechanisms for launching into the Describe VI feature. Seen here is the Tools menu.

LabVIEW Tools menu

The AI Advisor displays a static image of the VI, followed by a table summarizing its controls and indicators, and a plain-language description of what the VI does. This layout allows the engineer to quickly scan the structure, understand key parameters, and grasp the overall function without manually tracing wires or opening subVIs.

The user can ask follow up questions and even ask for help in how to extend the functionality.

Describe VI Wireframe

Result

Feedback showed that the AI advisor greatly accelerated engineers’ ability to interpret and repurpose measurements written by other engineers.

The new AI advisor improved the speed and confidence with which engineers could understand shared measurement code. In user feedback sessions, engineers reported that the AI-generated explanations, paired with the VI image and controls and indicators table, reduced the time needed to evaluate a VI. They also noted fewer dependencies on peers to clarify unfamiliar code, especially in time-sensitive validation work.

Describe VI Demo

The public announcement and demo of the AI Advisor at NI Connect 2025

My Role and Contributions

  • Led the design of the AI Advisor user experience and interface, transforming an engineer-built beta into a polished, user-focused experience aligned with IBM’s AI Design Principles.

  • Synthesized user feedback, AI capabilities, and technical constraints to create a clear UX vision that balanced empowering users with ethical AI use.

  • Worked cross-functionally with product managers, stakeholders, customers, and engineers to ensure the design was feasible, valuable, and well-integrated.

  • Advocated for user trust and transparency in AI explanations, helping establish design patterns that build confidence without overpromising.