Introduction

The rapid advancement and adoption of AI technologies, including large language models (LLMs) and chatbots, marks a new era in workplace and enterprise support by enabling the generation of human-like artifacts. Through a broad range of machine learning (ML) algorithms and deep neural networks, AI enables mission-critical functions such as pattern recognition, decision-making, and automation. To navigate this exciting albeit complex technological landscape, we introduce a human-centered AI (HCAI) solution map, which classifies AI solutions according to two dimensions: human-centered AI purpose, i.e., whether the AI solution is primarily intended to assist, augment, automate, or replace the human expert, and human-centered AI cognition, i.e., the type of cognitive function support such as sensing, decision-making, and action execution. Moreover, the HCAI solution map organizes AI solutions into roles ranging from assistants and tools to orchestration agents and automation agents. These roles vary in the way they support users and reflect different levels of human involvement during the process (i.e., in-the-loop, on-the-loop, or out-of-the-loop). Designing AI systems with human-centered qualities such as empathy, transparency, and explainability fosters positive user experience and trust. Characteristics such as context awareness, human-centeredness, and human-control, in turn, determine the effectiveness of AI solutions in adapting to situational needs and corresponding user interactions. Ultimately, by considering these human-centered dimensions, characteristics, and qualities when designing AI solutions, organizations can align the AI solution with the intended users’ needs and empower users and maintain human control while creating value for all stakeholders.