Agentic AI Framework Automates RTL Verification, Boosting Coverage.

Xin Xin, Jincheng Lou, Junhui Li, Jinglin Yan, Panda Xiao, Di Wu, Haixiao Li, Weicong Lu, Weijian Fan, Xinyu Qu, Yuxiang Zhao, Min Yu, Zhixiong Di, Yibo Lin· July 31, 2026 View original

Key takeaways

  • GoGoTB automates end-to-end RTL verification using an agentic AI framework.
  • It achieves high functional coverage by linking verification to specification requirements.
  • The system significantly reduces manual effort and potential for human error in IC design.
  • This represents a notable advancement in applying LLMs to complex engineering verification tasks.

Who benefits

SemiconductorAutomotiveAerospaceConsumer Electronics

Summary

GoGoTB is an agentic framework designed to automate the functional verification of integrated circuits, achieving high coverage rates without human intervention. It addresses challenges in existing LLM-based approaches by integrating an agentic execution control, an evolvable knowledge system, and specification-grounded coverage closure.

Functional verification is a critical and labor-intensive part of integrated circuit design, where undetected bugs can lead to costly hardware respins. Current AI methods for this task often struggle with maintaining context across different components and linking coverage directly to design specifications. GoGoTB introduces an agentic framework to overcome these limitations. This new system features an agentic execution layer that separates deterministic actions from LLM reasoning, a dynamic knowledge base for methodology and design expertise, and a coverage framework that ties every verification point back to a specific design requirement. This ensures that any remaining gaps are traceable and addressable. Evaluated on eight register transfer level (RTL) designs, GoGoTB successfully generated complete verification environments and achieved high levels of line, branch, toggle, and functional coverage, demonstrating a significant advancement over previous automated verification efforts.

Why it matters

This research offers a path to significantly reduce the time and cost associated with IC verification, improving product quality and accelerating time-to-market for hardware-dependent industries.

How to implement this in your domain

  1. 1Investigate GoGoTB's architecture for potential integration into existing verification flows.
  2. 2Pilot agentic verification tools on non-critical RTL modules to assess performance and identify integration challenges.
  3. 3Develop internal expertise in prompt engineering and agent configuration for specialized verification tasks.
  4. 4Establish clear metrics for evaluating the effectiveness and coverage completeness of AI-driven verification.

Original post by Xin Xin, Jincheng Lou, Junhui Li, Jinglin Yan, Panda Xiao, Di Wu, Haixiao Li, Weicong Lu, Weijian Fan, Xinyu Qu, Yuxiang Zhao, Min Yu, Zhixiong Di, Yibo Lin

"arXiv:2607.26181v1 Announce Type: new Abstract: Functional verification dominates integrated circuit (IC) front-end engineering effort, and a single missed bug that escapes to silicon can trigger a costly respin. Recent large language models (LLMs) offer new opportunities to auto…"

View on X

Originally posted by Xin Xin, Jincheng Lou, Junhui Li, Jinglin Yan, Panda Xiao, Di Wu, Haixiao Li, Weicong Lu, Weijian Fan, Xinyu Qu, Yuxiang Zhao, Min Yu, Zhixiong Di, Yibo Lin on X · view source

Want to go deeper?

Turn these trends into skills with Learnijoy's hands-on AI & tech courses.

Explore courses