Cisco Uses AI to Refactor 98 Percent of Legacy Code

The company is testing a two-person, AI-assisted approach to modernize its infrastructure over the next two years.

Updated on Oct. 30, 2026 in Artificial Intelligence

Isometric editorial illustration featuring a dense, structured lattice of geometric prisms representing complex digital infrastructure. — Cisco has initiated a 24-month project to modernize its aging software infrastructure by using artificial intelligence to refactor legacy code. AI Illustration. Upload story photo >

Live Poll

Do you trust artificial intelligence to manage and update critical software infrastructure?

Cisco is testing an artificial intelligence-driven approach to refactoring its legacy code base. The initiative relies on a two-person team to tackle the migration process over a 24-month period.

Why it matters

The company aims to modernize its aging software infrastructure by automating code replacement at scale. This project highlights a shift toward using AI tools to manage technical debt in large-scale enterprise environments.

Cisco is deploying a two-person team approach to overhaul 98% of its legacy code. The strategy begins with a four-week proof period to validate the efficacy of the AI-supported refactoring process.

The players

Cisco

A multinational corporation providing enterprise networking hardware, software, and telecommunications equipment.

EzDubs

A technology firm acquired by Cisco that specializes in AI-driven software development tools.

Padmanabhan Krishnamurthy

A 27-year-old software engineer and co-founder of EzDubs.

The details

The process uses artificial intelligence to scan and rewrite outdated code modules that represent the company's legacy stack. A two-person team manages these AI-assisted workflows to ensure functional consistency during the transition. The initial four-week proof period serves to measure the accuracy and speed of these automated refactors before scaling the project across the broader organization.

Nov. 25, 2025: Cisco completed the acquisition of EzDubs.
Next 24 months: Cisco aims to address 98% of legacy code.

The Tech Race

This project follows the 2025 acquisition of EzDubs, marking a direct attempt to integrate acquired AI talent into core engineering operations. It aligns with a broader industry trend of leveraging automated tools to retire decades of technical debt.

The changes will primarily affect Cisco's internal infrastructure and developer workflows rather than end-user hardware functionality. The four-week proof period will determine the feasibility of the proposed 24-month timeline for the codebase overhaul.

The takeaway

The project relies on a two-person team to demonstrate the feasibility of AI-driven code refactoring. Observers should track updates following the four-week proof period to see if the company expands the team size or adjusts the 24-month completion goal.