OpenAIOfficial ReleaseOfficial Source

1Password increases engineering productivity 21% with Codex

1Password engineers use Codex for features and internal tools while retaining production and security requirements. The useful signal is adoption inside a security-sensitive workflow.

OpenAI · Sep 8, 2026

Open Source Open Markdown Open JSON

Source Summary

Engineers at 1Password use **Codex** to build both **new features and internal tools**, with the work reaching production readiness under the company’s security policies.

Practical Implication

Builders introducing coding agents should keep production criteria and security controls in the workflow instead of treating agent output as exempt from normal engineering gates.

Agent-Ready Context

Engineers at 1Password use **Codex** to build both **new features and internal tools**, with the work reaching production readiness under the company’s security policies.

Builders introducing coding agents should keep production criteria and security controls in the workflow instead of treating agent output as exempt from normal engineering gates.

The supplied material does not define the productivity measurement, study design, time period, or specific security controls, so the size and generality of the reported benefit cannot be assessed.

Connected Context · Feed7 Judgment

This adds a security-sensitive production case to enterprise coding-agent adoption and confirms that agent-authored features and internal tools can remain subject to ordinary release controls. The reported 21% gain is more concrete than many adoption claims, but its missing methodology prevents comparison or use as a general productivity forecast.

Agentic SDLC at Uber — Uday Kiran Medisetty & Adam Huda, UberUber identifies governed access, isolated environments, validation, skills, and shared context as infrastructure behind scaled adoption, supplying implementation concerns left unspecified in the 1Password account.Which AI startups actually land enterprise contracts? — Brian Lewis, MillenniumThe buyer-side emphasis on security, controls, integration, and support reinforces the target’s implication that enterprise value depends on operational safeguards as well as agent capability.Government of Alberta uses Claude to find and fix cybersecurity vulnerabilities across government systemsAlberta’s human-gated patches reinforce the same principle that security-sensitive agent output remains subject to review, while demonstrating it in a large parallel-agent security audit.Australian Payments Plus moves faster with ChatGPT and CodexBoth are regulated or security-sensitive adoption cases that retain human or organizational controls, but neither supplies enough measurement detail for a clean comparison of claimed gains.

Context Map

toolscodingsecurity#coding-agents#enterprise#adoption

Uncertainty

The supplied material does not define the productivity measurement, study design, time period, or specific security controls, so the size and generality of the reported benefit cannot be assessed.