AI can now write code from plain English.
With OpenAI Codex, you describe what you want to build, and the AI writes the code for you. This is yet another leap forward by AI as it expands its capabilities.
Hello and welcome back to AI Six, a short-form daily podcast on how AI is changing marketing, communications, and digital work in six minutes or less.
What is Codex?
Codex is an AI-powered coding assistant developed by OpenAI. They describe it as:
“A cloud-based software engineering agent that can work on many tasks in parallel, powered by codex-1.”
It’s currently available to ChatGPT Pro, Team, and Enterprise users today, and will be available to Plus users soon.
Codex allows users to write prompts in plain English – no programming required – and generates working code in response. It’s designed to support tasks like building dashboards, analysing data, fixing bugs, writing documentation, and more.
Codex is integrated directly into ChatGPT, and it appears in the sidebar as an interactive tool where you can assign code-related tasks, review results, and iterate without switching platforms.
Core Capabilities
Codex supports over a dozen programming languages, but is most effective with Python.
Its capabilities include:
- Writing code from natural language prompts
- Debugging and explaining existing code
- Running tests and generating test output
- Writing documentation
- Producing complete pull requests for developer review
It also operates in isolated workspaces, meaning it can handle multiple tasks simultaneously. This enables users to run separate jobs in parallel – for example, analysing campaign performance, generating reports, and updating scripts at the same time.
OpenAI’s Codex can successfully complete about 37% of coding tasks on the first try, based on OpenAI’s own testing. If it’s given more chances—up to 100, it gets the right answer over 70% of the time. This shows Codex can help speed up routine coding tasks, but it still needs human support for more complex problems.
Safety and Governance
Codex runs in a locked-down, secure environment. It doesn’t have internet access and can only use approved tools, which helps prevent it from creating unsafe or harmful code. It’s also been trained to say no to requests that break the rules or could be dangerous.
Because of this, it’s safe to use with sensitive data, like client campaigns or internal reports.
Use Case – Digital Marketing Campaign Analysis
Codex is particularly useful for marketing professionals working with campaign data.
For example, if you have a CSV file with campaign metrics – impressions, clicks, conversions, and spend – across multiple channels, you can prompt Codex to:
- Identify top-performing platforms based on conversion rates and ROI
- Segment audiences by engagement
- Detect seasonal trends
- Generate bar or line charts for reporting
- Provide actionable insights and next steps
You can also ask Codex to return the Python code used for analysis, so the process can be automated in the future.
This workflow can reduce dependency on developers or data teams and accelerate strategic decision-making.
Writing Better Prompts for Codex
To get the most out of Codex, follow these best practices when writing prompts:
- Start with a clear objective – e.g., compare campaign ROI by channel over the last quarter.
- Be specific about your data – list the columns, data types, and timeframes.
- Include context – mention your business goals or target audience.
- Ask for insights – not just summaries, but recommendations or visual outputs.
- Specify the format – whether you want bullet points, charts, or code.
- Refine iteratively – adjust the prompt if the initial result isn’t complete.
A well-written prompt can save time and produce much more accurate results.
Codex gives marketers, analysts, and non-developers access to real coding power without needing to write scripts from scratch.
By combining natural language prompts with secure execution and smart outputs, it reduces friction, accelerates workflows, and enables deeper data analysis with minimal overhead.
If you’re already using ChatGPT Pro, Codex is now available and worth exploring.
About the Podcast Host
Joanne Sweeney is a specialist in public-sector marketing and digital communications. She’s the author of Public Sector Marketing Pro and the creator of the Professional Diploma in AI for Public Sector Marketing. With years of experience in training, consulting, and advocating for digital transformation, Joanne provides insight and guidance for government agencies navigating the digital landscape. She is also the host of the award-winning Public Sector Marketing Show podcast.



















