- Title: AI Task Manager & Operations Agent
- Category: Internal Operations & Task Automation
- Context: Small business operations (bakery production and delivery). The owner handles dozens of daily micro-tasks (ordering ingredients, contacting suppliers, checking packaging stocks).
- Problem: Tasks were written down manually in chat apps or notebooks, IDs were lost, incomplete tasks were untracked, and updating task statuses required manual table entries.
- Solution: Designed an autonomous conversational AI Agent with short-term memory and 3 isolated tool subworkflows. The agent interacts via natural language, executes actions, and maintains structured data inside n8n Data Tables.
- Process:
- Prompt Design: Configured a system prompt with strict action routing rules (add, list, or complete by ID) and anti-hallucination constraints.
- Architecture: Developed a 4-workflow architecture (1 Main Agent + 3 Subworkflow Tools).
- API & Storage: Integrated n8n Data Tables (
id, title, status) and LLM APIs for natural language intent processing.
- Results:
- ⏱️ 1.5+ hours saved daily: The owner sends quick text/voice notes instead of managing spreadsheets.
- 🎯 0 lost tasks: Instant task logging and status retrieval in seconds.
- ⚡ Response Speed: Instant execution under 1.5 seconds.
- Tech Stack: n8n (Chat Trigger, AI Agent Node, Subworkflows), OpenRouter / OpenAI API, n8n Data Tables, JavaScript.
- Visuals:
