Introduction

AI for software development is here, practical, and professionalized. Developers are writing code with it, testing systems with it, and learning how it fits into building applications.

A recent study of 147 professional developers sheds light on how that work is taking shape. The research reveals distinct, stable patterns in how developers use AI, how they perceive its impact on productivity and quality, and how organizational context shapes those experiences. Our research identifies three distinct archetypes that describe how AI use is structured inside teams: Enthusiasts, Pragmatists, and Cautious developers.

Three Developer Archetypes

The three empirically distinct developer archetypes occupy stages in a technology diffusion process. Enthusiasts are similar to early adopters, Pragmatists are the early majorities, while the Cautious are the late majority and some laggards.

Enthusiasts: Proving AI Tools Work

Enthusiasts are the vanguard of AI use in software development and are at the forefront of a broader trend: nearly 75% of all developers surveyed report an improvement in perceived code quality attributable to AI coding tools. They score highest on usage breadth, optimism, strategic outlook, and satisfaction with current code quality. They report strong productivity gains and view the current generation of AI tools as successful. They manifest the “Virtuous Adoption Cycle”: high usage leads to high perceived productivity, which in turn reinforces their belief that AI is actively improving their work.

What most clearly differentiates Enthusiasts is policy support. Nearly 60% report explicit AI coding policies in their organizations—the highest of any group. In these environments, AI use is legitimate, visible, and normalized. Policy functions less as restriction and more as a marker of organizational maturity, formalizing practices that have already demonstrated value.

Significantly, they adopt ahead of endorsement, essentially proving the “proof of concept” for the rest of the company.

Policy Shapes Archetypes

One of the most important insights from the research is how policy actually functions in AI adoption. Our analysis shows that intent is driven by frequency of use and ease of integration. But policy determines who gets the opportunity to build that experience in the first place.

For Enthusiasts, policy marks maturity—codifying success after the fact. For Cautious developers, policy acts as a gate, legitimizing experimentation and lowering perceived risk. This dual role explains why policy clusters so strongly with Enthusiasts and is nearly absent among the Cautious.

Where policy is explicit and enabling, Enthusiast behavior emerges and persists. Where policy is partial, Pragmatist behavior dominates. Where policy is absent, Cautious behavior is reinforced.  Moreover, AI use in coding is already widespread. AI use in testing lags and represents a huge opportunity to gain further productivity and quality improvements.