The Persona You're Ignoring Will Break Your Product

Product teams have long focused on creating detailed personas to understand their users better. These personas typically include archetypes like the Stressed Admin, the Power User, and the Busy Executive. However, with the rise of AI, these personas have been left behind. AI has been treated as a mere feature rather than a genuine user to design for. This strategic oversight leads to superficial chat bubbles and abandoned features, as users quickly grow tired of them.

Moe Ali's teachings emphasize the importance of focusing on contrast when designing products. AI is not just another feature to add; it requires a whole new approach to product design. Products must stop treating AI as an afterthought and start designing for it as the most demanding user on their platform. Human users often compensate for poor design, but the Agent, an AI user, lacks the intuition to do so effectively. AI cannot read visual cues or understand unwritten rules, leading to failures and incorrect actions.

To design for the screenless persona, product managers must rethink their acceptance criteria. Instead of focusing on clicks, they should design for intent. The agent cannot perform tasks that require human intuition, such as understanding contextual cues or following unwritten rules. Products must be designed to recognize and respond to the agent's needs, using verbs rather than simulated button presses.

Context is crucial; if the product relies on human knowledge, it should be explicitly defined within the product's logic.

Granular trust levels are essential, as autonomy should not be a simple master switch. Actions that affect the real world must require human confirmation. The agent's success rate may appear high, but it will quickly lead to errors and user dissatisfaction. Product teams must reconsider core features that work seamlessly for humans but are illegible to AI. The new primary metric should be the number of tasks completed without human correction, which often reveals the limitations of AI readiness.

Incorporating the Agent into the persona wall is crucial. Product teams must understand its needs, constraints, and goals. Until every new feature spec asks, "Can a screenless agent complete this end-to-end?" the product is not truly AI-ready. The Agent's frustrations include undocumented rules, state that exists only on the screen, and success responses that are not accurate. By focusing on how AI is described rather than how well it works, product teams can ensure AI readiness and avoid silent churn mid-sentence.