Cisco claims to have saved more than 1,500 engineering hours per month from using OpenAI’s Codex AI coding agent platform.

The networking giant has been working with the ChatGPT maker for the past year, with Cisco CPO Jeetu Patel revealing it was “exploring” Codex, testing the tool’s potential as it sought a 10-times increase in productivity from “the human workforce.”

But a newly published blog post details Cisco’s broad deployment of Codex, with its software engineers using it to reduce project build times by as much as 20%, while also reporting up to a 15-times increase in defect resolution throughput.

“Codex has become a meaningful part of how we think about AI-assisted development and operations going forward,” Brad Murphy, a VP leading Cisco’s Splunk engineering team, explained.

Cisco’s shift from evaluating AI technologies to actively using them to make offerings production-ready sees Codex integrated directly into production engineering workflows. The coding tool is exposed to multirepository systems, C/C++-heavy codebases, and the security, compliance, and governance requirements.

The company’s engineers were able to give feedback directly to OpenAI on how the tool worked in actual production environments to better shape workflows.

The result of Codex’s integration has seen Cisco’s Splunk teams drastically reduce the time it takes to perform migrations from weeks into days, a change the blog claims then allows its engineers to “concentrate on judgment-heavy decisions.”

“The biggest gains came when we stopped thinking about Codex as a tool and started treating it as part of the team,” Ryan Brady, a principal engineer in Cisco's Splunk group, added. “We use Codex to generate and follow a plan document, allowing the reviewing team to more easily understand both the process and the code generated.”

Cisco has already offered glimpses into how AI is augmenting its external-facing offerings.

In December, it showcased a security-centric agentic AI offering capable of automating certain security operations center (SOC) use cases, such as incident investigation. While it’s also exploring ways for customers to use AI to accelerate data extraction from across their networks.

But internally, there were suggestions Cisco was looking at AI tools as a means to streamline its workforce, a position the company has sought to quash.

“As an AI infrastructure company, we must attract the best developers in the world to Cisco. That can only happen if they can do their life’s best work with us,” Patel wrote last May. “We want Cisco to be a company where engineers have the best tooling in the world and are empowered to fundamentally reimagine how they write code.”

CEO Chuck Robbins went one step further in August, declaring Cisco would not be cutting jobs in favor of AI – though some may argue his words rang hollow when days later, news broke that some 150 employees were to be axed as part of a new round of layoffs.

AI was among the reasons Cisco cited for cutting 7% of its workforce in August 2024, a decision CFO Scott Herren suggested at the time was more akin to “a reallocation versus a headcount savings.”

Those cuts were preceded by a 5% reduction of Cisco’s global workforce that February, a restructuring move the vendor suggested at the time would maximize shareholder returns.