The Chinese writing system has enjoyed the heritage commonly known as “Writing and Painting Are One”. During multiple major transformations in the past three thousand years, the pictorial features like long curves have been replaced by short and straight essential strokes and composite strokes with essential strokes as their parts. Knowledge about the original design patterns of Chinese characters, which convey meaning and sound, is incomplete and currently underutilized in teaching. This helps build the impression that Chinese is the hardest to learn. Although comprehension of character composition and its consequences on a character’s meaning has seen continued improvements from ChatGPT 3.5 to current versions like DeepSeek, a deficiency of the original “design” becomes a bottleneck for fully exploit the potentials of computational intelligence of teaching Chinese to adult non-heritage students. This paper presents data and knowledge models including stroke codification that can be used to build an AI teaching system with rule-based inferencing to use a “storytelling” style to explain meanings of the characters exploiting pictorial hints as embedded in ancient scripts. Personalized teaching materials and game-like exercises can also be generated.