Nielsen Norman Group Examines Role Of UX In Agentic Software Factories

The Nielsen Norman Group reported on the potential risks of adopting AI-native software development lifecycles. These lifecycles reduce UX to AI context, which may lead to a loss of user insight. This is a concern because UX teams play a crucial role in discovering user needs and shaping what gets built.

In recent months, several major companies, including Anthropic, Google, and OpenAI, have published their visions for AI-native software development lifecycles. While these visions differ in details, they all seem to sideline UX, reducing it to behind-the-scenes context that guides AI generation. However, UX offers more value than just providing context.

The traditional software development lifecycle consists of six stages: plan, design, build, test, deploy, and maintain. Early versions of this lifecycle emerged in the 1960s, and it has since evolved with the Agile movement. Despite its evolution, the lifecycle has primarily been described from an engineering perspective, with UX teams fighting to defend their place in the process.

The AI-native software development lifecycle, as described by Anthropic and others, retains the six traditional stages but with AI as the primary executor. Humans monitor quality along the way, and the stages revolve around a chain of version-controlled artifacts generated by AI agents. The goal is to maximize productivity gains by adapting the other stages to keep pace with AI agents.

A key aspect of the AI-native lifecycle is the use of context to guide the AI's work. This context includes curated instructions that tell the AI how the organization does things and constrain its work. However, this approach may not be sufficient, as UX teams bring more value to the table than just providing context.

The Nielsen Norman Group argues that UX teams must discover user needs, shape what gets built, and evaluate what gets built before and after launch. By reducing UX to just context, organizations may risk losing the very user insight their AI context depends on. This could have significant consequences for the success of their products.

To address these concerns, UX teams can take several steps. They can start by evaluating the AI-native software development lifecycle and identifying areas where they can add value. They can also work to strengthen their shared judgment about what to build and ensure that user needs are being met.