Why it matters
AI coding assistants are now common in software teams. Leaders want to capture productivity gains while maintaining quality, security and developer skills.
Coding assistants change how developers write, review and test code. Organisations are learning where they help most, how to maintain code quality and how to develop junior engineers' skills.
By the numbers
73%of developers who use AI coding assistants or agents use them every day 1
66%of AI users work with coding agents; 63% use chatbots 1
93%say AI output needs source attribution before they can trust it 1
All respondents
Daily AI users
Source: Stack Overflow, The results of the 2026 Developer Survey are here, 6 October 2026
What's changing
- Assistants help with code completion, testing and documentation.
- Organisations are setting policies on code provenance and licensing.
- Productivity measures are moving towards delivery outcomes.
How leaders are responding
- Rolling out assistants with training and clear usage guidelines.
- Reviewing AI-generated code with the same rigour as other code.
- Measuring outcomes such as lead time and change failure rate.
What good looks like
- Code review and testing standards that apply equally to AI-written code
- Productivity measured on delivery outcomes, not lines of code
- Approved tools with clear rules on code and data sharing
What to watch
- Agentic tools that complete multi-step development tasks.
- Security scanning of AI-generated code.
- Changes to engineering hiring and training.
Where to start
Run a structured trial with a few teams, measuring delivery and quality outcomes before a wider rollout.
Questions for your leadership team
- How are our developers using assistants today?
- Which controls do we need for security and licensing?
- How will we measure the impact?
