“I want to build an AI that suggests restaurant menus from inventory and weather”
AirREGI / Smaregi-integrated SaaS "WeatherStock AI" (PoC → monthly subscription)
The value hypothesis of linking inventory (POS/menu) and weather has academic support, and POS APIs make implementation realistic. First launch a Google Sheets-based PoC integrated with AirREGI/Smaregi within 7 days, run a paid trial at one Tokyo store, quantify results (sales difference and waste reduction), then proceed to a full SaaS product.
Steps to start today
- 1Create a Python script to retrieve sales and menu data through the AirREGI / Smaregi APIsSteps: review the AirREGI FAQ and Smaregi API documentation, then create a Python script that uses the store's API key to retrieve daily sales CSV data and the menu master. Deadline: 3 business days. Tools: Python, requests.
- 2Implement a weather-scoring module with the OpenWeather APISteps: create a function that calculates a weather score (0–100) from the temperature and precipitation probability for today through three days ahead using OpenWeather's free plan. Deadline: 2 days. Cost: run within the free API tier. Tools: OpenWeather API, Node/Python.
- 3Build a PoC dashboard with Google Sheets and Apps ScriptSteps: load the sales/menu data and weather score into Google Sheets, then use Apps Script to display the top three menu recommendations for the day. Deadline: 4 days. Purpose: make it viewable by store staff.
- 4Implement a simple rule-based recommendation algorithmSteps: score estimated inventory as the average units sold over the past 7 days × a safety-stock factor, and weather sensitivity as the weather score × each product's sensitivity. Recommend the top three products. Validation metric: measure the sales difference ±% during the trial. Deadline: 5 days. Tools: Python/Sheets.
- 5Interview the test store and obtain implementation consent (prepare a contract template)Steps: present data-access feasibility, expectations, and two-week trial terms (¥15,000 implementation fee plus ¥4,980 monthly fee) to one independent izakaya in Tokyo and obtain consent. Prepare a contract template. Deadline: 7 days. Tools: email, in person.
Reference and validation data
- FACTMajor POS systems such as AirREGI support API data integration, making sales and menu data practically retrievable.How to use API data integration with various systems – AirREGI FAQ
- FACTAcademic and applied research shows that weather (temperature, precipitation, and other factors) affects menu-level restaurant sales.The effect of weather factors on restaurant sales
- FACTOther POS systems such as Smaregi also publish platform APIs, so POS integration can be implemented across multiple providers.Smaregi Platform API Reference
Smallest deliverable to build first
Build a PoC dashboard with Google Sheets and Apps Script: import AirREGI/Smaregi CSV data (or use an API), combine it with a three-day weather score from the OpenWeather API, and display three rule-based menu recommendations for the day (deadline: 7 days).
First customer to target
Propose a two-week paid trial using real data to an independently operated izakaya in Tokyo (around 20 seats and using a POS such as AirREGI or Smaregi). Goal: secure one test store and finish negotiations within 2 weeks.
Concrete monetization path
Recruit early adopters at an implementation fee of ¥15,000 plus a monthly fee of ¥4,980 (earn initial revenue from implementation support and start billing within one month).
Biggest pitfall or risk
The store may not consent to API access, or POS fields may be missing and prevent accurate estimates. Confirm permissions and data fields before implementation.