Task Modeling
During exploration, task modeling utilizes magic-enhancing-magic to gain an understanding of task-related knowledge, obtain preliminary task requirements, workflow, input/output data and constraints, and initially align task decomposition with model capabilities, while also exploring different prompts to gauge their effectiveness.
Such an initial task model provides the foundation for system design and will be continuously elaborated during the AI chain design and construction process, and further evolved driven by the end-users' feedback after deployment.