Best Practices

Best Practices for Interacting with Finnex AI

Follow these guidelines to ensure smooth, efficient, and cost-effective use of the FINNEX AI Assistant.

  • 🧩 When Interacting with Excel

    • Always specify the range (e.g., Sheet1!B2:D10) when asking AI to read or write data.

    • If no range is provided, AI will act on the active cell by default.

    • Use “My Excel” in your query for better recognition.

    📎 When Interacting with Attachments

    • Use the Upload button to attach documents (PDF, PNG, JPG, TXT, etc.).

    • Multiple files can be uploaded, but the combined size must stay within 50,000 tokens (~40k words).

    • Refer to files as “Attached file” when asking AI to extract data or answer questions.

    🗃️ When Interacting with the Database

    • Ensure uploaded data is clean and well-structured.

    • Always include a Unique Identifier (Primary Key) when updating records.

    • Use “My database” or “My DB” in your prompts for accurate database handling.

Best Practices for Efficiency & Cost Control

These practices help reduce token usage and optimise performance:

  • 🔄 Start Fresh in a New Sheet

    • AI outputs overwrite Excel content. Use a new sheet to avoid accidental data loss.

    ⚠️ Avoid Sending Large Datasets to AI

    • GPT models are not optimised for processing large datasets directly.

    Best Practice:
    Upload large data sets to the Database tab, then ask AI to analyse data from there.

    🧹 Clean Up Chat Threads Regularly

    • Older threads add context, increasing token usage and slowing responses.

    Best Practice:
    Start a new thread for new topics. Use AI to summarise previous threads if needed before continuing.

    🧠 Be Specific & Concise

    • AI performs best with clear and to-the-point instructions.

    Best Practice:
    Avoid vague or overly long prompts. Treat the AI like a skilled assistant who doesn’t know your business yet.

    🧰 Use Built-In Features Instead of Repetitive Prompts

    • Use the Excel-Database integration, automation library, or predefined templates where possible.

    Best Practice:
    Let built-in tools handle repetitive steps to reduce token consumption.

    🧼 Preprocess and Filter Data First

    • Pre-filter your data in Excel before passing it to AI for analysis.

    Best Practice:
    Limit the data scope to only what's relevant. This minimizes token usage and improves performance.

    📊 Monitor and Manage Your Credits

    • Credits = Tokens. Watching your usage helps you control costs.

    Best Practice:
    Use the dashboard to monitor token consumption and plan accordingly.

    📅 Plan Your Usage in Advance

    • Anticipate your AI needs and buy credits accordingly.

    Best Practice:
    Buy credits in bulk when discounted or during high-usage periods. Choose the right plan based on your patterns.

    🤖 Use Automation (Enterprise Users Only)

    • The Library interface allows scripting of common tasks to save time and tokens.

    Best Practice:
    Automate routine tasks to reduce repeated AI queries.

    🔧 Review and Refine Scripts Regularly (Enterprise Users Only)

    • Outdated scripts may be inefficient or error-prone.

    Best Practice:
    Regularly review and update your automation scripts for improved accuracy and token efficiency.