Finance Board Pack AI Automation

This is an automation that turns monthly Google Sheets actuals into MD&A commentary (through OpenAI API), writes to Notion, and notifies Slack — triggered via Webhook.

---

1. Repository structure

finance-board-pack-ai-automation/
├─ README.md
├─ LICENSE
├─ workflow.json                 # High-level list of Make modules
├─ data/
│   └─ board_model_example.csv   # Example “Board_Model” Google Sheet
├─ demo_output/
│   ├─ sample_notion_row.md
│   └─ sample_slack_message.md
├─ docs/
│   └─ architecture.md           # Text architecture diagram
├─ screenshots/
│   ├─ Google_Sheets_Actuals.png
│   ├─ Make_Scenario_Overview.png
│   └─ Notion_DB_Structure.png

---

2. Screenshot gallery

Images are stored in screenshots/ and linked relative to README.

Google Sheet structure

Make.com automation scenario

Notion database structure

---

3. How the Make scenario works (high level)

1. Webhooks – Custom webhook

Entry point. Opening the webhook URL (or calling it from another tool) triggers the full run.

2. Tools – Set Variable (BASE_CCY=…)

Stores the base reporting currency (e.g., GBP).

3. HTTP – Make a request (FX API)

Calls your FX provider (e.g., ExchangeRate.host / ECB demo).
Response is mapped to a variable and later stored in Notion as raw JSON.

4. Tools – Set variable (FX_RATES)

Keeps the selected data rates object handy for later use.

5. Google Sheets – Get Range Values

Reads Board_Model → ‘actuals’ tab
Headers: entity, currency, period_yyyy_mm, metric, amount.

6. JSON – Create JSON (Actuals_Structure)

Wraps the rows into a JSON array with fixed schema:
entity, currency, period_yyyy_mm, metric, amount.

7. OpenAI – Generate a completion

Model: gpt-4o (or later).
Prompt: describes the group performance based on the JSON rows, in professional MD&A language.

8. Notion – Create a Database Item (Legacy)

Database: Board Pack Run Log
Fields typically mapped:

• Run ID – timestamp or static prefix + period
• Period – from the sheet (e.g., 2025-07)
• Entities – list of entities found
• Base Currency – from JSON
• FX Rates (JSON) – raw JSON string from HTTP step
• MD&A Summary – trimmed OpenAI completion text
• Board Pack PPTX Link – optional future extension
• Requested By – webhook source / caller

9. Slack – Send a Message (optional)

Posts a short summary plus a link to the Notion row / Sheets / Drive.

For a visual overview, see docs/architecture.md.

---

4. How to reuse this for other companies

Step 1 – Duplicate items

Duplicate the:

• Google Sheet
• Notion database
• Periods / months
• Metrics / amounts structure

Step 2 – Point the Google Sheets module to the new spreadsheet

Keep the same header structure.

Step 3 – Point the Notion module to the new Board Pack Run Log database for that client.

Step 4 – Modify the OpenAI prompt

Optionally adapt to each client’s tone, KPIs, sector, metrics.

Step 5 – Everything else stays the same

The scenario architecture remains identical, enabling scalable reuse across your fractional-CFO portfolio.

---

5. Demo artefacts

Located inside /data and /demo_output:

• data/board_model_example.csv — ready to import into Google Sheets as your “Board_Model → actuals” tab
• demo_output/sample_notion_row.md — typical Notion row output
• demo_output/sample_slack_message.md — Slack summary preview

---

6. License

This project is released under the MIT License. Use it freely as a template for your finance-automation portfolio.

---