Overview

I built a no-code AI agent for a small flower shop to explore how conversational agents can move beyond answering questions and start completing real business tasks. The prototype handles a common customer-service workflow: answering store questions, checking flower availability, collecting order details, updating inventory, and logging confirmed orders into Google Sheets.

Why It Matters

This prototype shows how small businesses could use agentic workflows to automate repetitive service tasks without building custom software. It could also extend customer support beyond normal business hours, allowing customers to ask questions, check availability, and submit order details even when the owner is unavailable.

Key Learnings

1. AI agents become more useful when they can take action, not just answer questions.
2. Google Sheets can work as a simple backend for early AI automation prototypes.
3. Make is effective for connecting conversational inputs to real business workflows.
4. Inventory checks need clear fallback logic for missing items or insufficient stock.
5. ChatGPT was useful as a troubleshooting partner while building flows in Relevance AI and Make.