AI Assistants for Electronic Design Automation

Keysight Technologies has introduced AI-powered Chat and Copilot assistants for its electronic design automation environment, enabling natural language interaction with design and simulation tools while maintaining enterprise-grade security controls.

Addressing complexity in modern EDA workflows
Electronic design teams increasingly work across RF, microwave, and high-speed digital domains, often within tightly coupled simulation and verification environments. These workflows are typically characterized by steep learning curves, fragmented data handling, and manual task repetition, all of which can slow development and increase the risk of errors.

Keysight’s new AI assistants are designed to reduce these constraints by embedding domain-specific language models directly into the EDA workflow. The goal is to accelerate routine operations and onboarding without changing established design processes or exposing sensitive design data.

Two assistants with distinct roles
The release introduces two complementary assistants that serve different stages of the design process. The Learning Assistant, referred to as Chat, provides conversational guidance on tool usage and application-specific questions. It is intended to support faster ramp-up for new users and quicker access to contextual information for experienced engineers.

The Tool Assistant, referred to as Copilot, operates more actively within electromagnetic simulation environments. It can execute commands and automate repetitive steps inside RF and signal integrity simulation flows, reducing the need for manual menu navigation while keeping the engineer in control of all actions.

Domain-specific AI training
Unlike general-purpose AI tools, the assistants are trained on electronic design–specific content derived from Keysight’s EDA expertise. This specialization is intended to improve the relevance and accuracy of responses for RF and high-speed design tasks and reduce the likelihood of misleading or non-applicable suggestions.

By focusing on a defined problem domain, the assistants are positioned as productivity tools rather than autonomous design agents, supporting decision-making without replacing established engineering judgment.

Security and deployment model
A central design requirement for the assistants is data control. Keysight has implemented the AI capabilities for on-premises deployment, ensuring that customer design data remains within the organization and does not leave local IT environments. This approach supports compliance with internal security policies and regulatory requirements common in aerospace, defense, and telecommunications sectors.

The assistants are deployed as a separate add-on to the main EDA software, giving IT teams control over installation, configuration, and permitted toolsets. Supported deployment options include containerized environments such as Kubernetes and Red Hat OpenShift running on commercially available GPU hardware.

Impact on EDA productivity
By enabling natural language interaction with design and simulation tools, the assistants reduce the need to navigate complex interfaces for routine tasks. This can shorten design cycles, lower training overhead, and help teams manage resource constraints more effectively.

Within the broader electronic design automation landscape, the introduction of secure, domain-trained AI assistants reflects a shift toward augmenting engineer productivity rather than automating design decisions outright. As EDA workflows continue to grow in complexity, such tools are increasingly positioned as a way to maintain development velocity without compromising security or design integrity.