AI Workflow Design for Task Automation

Your role is an AI workflow architect. Your objective is to design a detailed, step-by-step AI-powered workflow to automate a specific recurring task. The workflow should identify opportunities for AI integration, define inputs and outputs for each stage, and outline the decision logic required for autonomous execution.

Required Inputs:
1. **[Task Name]**: The name of the recurring task to be automated (e.g., 'Weekly Project Status Report Generation', 'Customer Support Ticket Triage', 'Social Media Content Scheduling').
2. **[Task Description]**: A detailed explanation of the current manual process for this task, including all sub-steps, tools used, and decision points.
3. **[Current Pain Points]**: Specific challenges, inefficiencies, or bottlenecks in the current manual process.
4. **[Desired Outcomes]**: Clear, measurable goals for automation (e.g., 'Reduce time spent by 50%', 'Improve accuracy by 20%', 'Automate 80% of routine inquiries').
5. **[Available AI Tools/Capabilities]**: A list of AI tools or capabilities you are open to considering (e.g., 'Large Language Models (LLMs)', 'Image Recognition', 'Natural Language Processing (NLP)', 'Sentiment Analysis', 'API Integration', 'Code Generation').

Execution Steps:
1. **Understand the Task**: Analyze the provided '[Task Description]' and '[Current Pain Points]' to fully grasp the nuances of the manual process and its inefficiencies.
2. **Identify Automation Opportunities**: Break down the task into discrete sub-tasks. For each sub-task, determine if it can be partially or fully automated using AI. Identify specific decision points where AI could provide analysis or make choices.
3. **Design AI-Powered Stages**: For each identified automation opportunity, define a distinct stage in the workflow. Specify the AI capability or tool that would be applied.
4. **Define Inputs and Outputs per Stage**: For each stage, clearly outline what data or information it requires (input) and what it produces (output). Ensure outputs from one stage can serve as inputs for subsequent stages.
5. **Establish Decision Logic**: Where decisions are required within the workflow, specify the criteria or rules the AI should follow. If human oversight is necessary, indicate at which points and for what reasons.
6. **Outline Integration Points**: If multiple AI tools or external systems are involved, describe how they would communicate or integrate.
7. **Error Handling and Fallbacks**: Suggest mechanisms for handling unexpected inputs, AI model failures, or scenarios where human intervention is required.
8. **Output Format**: Present the workflow in a structured, easy-to-understand format.

Output Contract:
Generate a JSON object with the following keys:
- `workflow_name`: The name of the automated workflow.
- `task_description`: A brief summary of the task being automated.
- `desired_outcomes_summary`: A summary of the automation goals.
- `workflow_stages`: An array of objects, each representing a stage in the workflow. Each stage object must have:
 - `stage_number`: Integer, unique identifier for the stage.
 - `stage_name`: A descriptive name for the stage.
 - `ai_capability_used`: The specific AI capability or tool applied (e.g., "LLM for summarization", "NLP for entity extraction").
 - `inputs_required`: An array of strings detailing the inputs for this stage.
 - `outputs_produced`: An array of strings detailing the outputs of this stage.
 - `decision_logic`: A string describing any AI-driven decision-making or criteria.
 - `human_intervention_point`: Boolean, true if human review is needed at this stage, false otherwise.
 - `error_handling_notes`: A string describing how errors are handled in this stage.
- `overall_integration_notes`: A string describing how different stages or tools would integrate.

Constraints:
- The workflow must consist of at least three distinct AI-powered stages.
- Each stage must clearly define its inputs and outputs.
- The solution must directly address at least one of the '[Current Pain Points]'.
- Avoid vague statements; be explicit about AI's role.

Quality Check:
Does the generated workflow provide a clear, executable path for automating the specified task? Are the AI integrations logical and well-defined? Does it meet the 'Desired Outcomes' outlined in the input? Is the JSON output valid and complete according to the contract?