Acne is an extensively occurring skin condition that can result in severe repercussions if left unaddressed or exacerbated. Traditional acne diagnosis, performed manually by dermatologists, lacks standardization, making accurate severity evaluation challenging - yet such evaluation is essential for effective treatment planning. Currently, there are approximately 25 different scaling systems in use for evaluating the severity of acne, but a global standard for grading has not been established due to the unique advantages and disadvantages of each method. As a result, this complexity can extend the duration of the conventional acne diagnosis process, which dermatologists conduct manually, potentially leading to subjective assessments of its severity. Therefore, this paper presents ACNE84-Alpha, an AI-based system developed using the YOLOv8 technology, to address these challenges. ACNE84-Alpha comprises two components: an acne detection model and an acne severity grading model. The acne detection model, named ACNE8M, can accurately identify seven primary and secondary acne lesion types, one severe clinical form, and four additional differential diagnoses. The acne severity grading model, trained on the ACNE04 dataset of approximately 1400 images, reliably evaluates acne severity. Extensive evaluations show that ACNE84-Alpha achieved outstanding performance, with the acne detection model achieving a mAP score of 0.69 across 12 classes and the severity grading model attaining a mean accuracy of 0.732. These remarkable results surpass existing methods and highlight the potential of ACNE84-Alpha to revolutionize acne diagnosis. By providing accurate acne detection, differential diagnosis, and severity grading, ACNE84-Alpha offers a comprehensive solution to support dermatologists and empower patients to understand their condition better, ultimately leading to improved treatment outcomes and enhanced quality of life.