Developer confidence in AI has increased over the last year, according to our 2026 Developer Survey. This is great news from an adoption standpoint. However, this also comes with some caveats and opportunities for improvement.
Practitioners tend to lean away from using AI when it comes to tasks that require more subjective judgment and human communication. This is entirely understandable. If a developer uses incorrect information, it can have big consequences. They want to do their job well, and that means being careful with the data they use. The possibility of incorrect AI outputs make it hard for teams to move from individual use to scaling agentic workflows across teams.
The 2026 Developer Survey released in October surveyed over 30K technologists across 169 countries with 103 questions in five categories: work, AI, technology, community, and knowledge. In this guide, we will uncover some of the key data from this survey around:
- What unreliable AI really costs
- Why the fix is providing better context
- Why the true AI return on investment is outcomes
Key takeaways:
- Unreliable AI outputs have a hidden tax on businesses.
- A lack of foundational context is a core problem for teams using AI. Practitioners have more confidence in answers from verifiable sources.
- Outcomes are the true value of AI integration that leaders should focus on, and better outcomes start with shared organizational knowledge that can be verified.
Each section concludes with practical guidance for every key buyer or decision maker from CIOs and Engineering VPs to CISOs and FinOps leaders. The goal is to help improve the practitioner experience with AI integration, allowing teams to use the tools they prefer, connect with the subject matter experts, and set goals focused on business outcomes.