Objective:

To leverage generative AI tools for developing an intelligent, user-friendly, and compliant utility that helps users determine the applicability of Reverse Charge Mechanism (RCM) under GST for service-based transactions.

Step 1: Initial Research Using ChatGPT

An initial search for RCM service transactions was made using ChatGPT. However, the list retrieved was found to be either outdated or incomplete, lacking the desired format and critical compliance parameters such as notification references and dates of applicability.

Step 2: Leveraging CA GPT – Indirect Taxes

To obtain a structured and compliance-aligned data set, the query was shifted to CA GPT – Indirect Taxes with the following prompt:

Prompt Used:

“Give me list of RCM transaction in services in the format:

S.No. | Description | Type of supplier - Registered/Unregistered | Type of recipient - Registered/Unregistered | Date of Applicability | Notification Number

(Provide in tabular form)”

Output:

A comprehensive table was generated outlining major RCM transactions in services, compliant with GST law for FY 2024-25. The table included:

1. Description of service
2. Supplier and recipient types
3. Date of applicability
4. Relevant GST notification numbers

This output became the base dataset for building further AI-driven logic.

Step 3: PDF Conversion

To ensure compatibility across various AI platforms, the structured table was converted into a PDF document, as many AI tools perform better when fed structured documents in this format.

Step 4: Integration with NotebookLM

The PDF was uploaded to NotebookLM, an AI tool designed for summarization and prompt creation. However, efforts to generate a refined prompt for a code generator AI (like Llamacoder) did not yield appropriate or technically viable results, even after multiple iterations.

Step 5: Code Generation with Claude.ai

The next phase involved Claude.ai—a powerful AI tool for code generation and app development. A simple prompt was issued by attaching the earlier generated PDF.

Prompt Used:

“By implementing these checks, create a code to determine the applicability of RCM based on the provided information. Ask the user for information using dropdowns.”

Output:

Claude.ai successfully generated the first version (V1) of an interactive tool where the user could:

1. Select the service type via dropdown
2. Choose the supplier type
3. Select the recipient type

The tool would then:

1. Inform if RCM was applicable
2. Display date of applicability
3. Cite the relevant GST notification

Step 6: Enhancement – Date-Based Applicability Logic (Version 2)

The tool was upgraded to include a “Date of Supply” input field.

Enhancements Made in V2:

1. Added a date comparison logic to verify if the supply was made on or after the applicable date.
2. Included a detailed write-up section below the results explaining:
3. Description of the RCM scenario
4. Legal references and dates
5. Whether RCM is applicable and why
6. Compliance steps if RCM applies

Special cases such as withdrawn services (e.g., Ocean Freight under CIF) were also integrated into the logic.

Step 7: Enhancement – Corrective Suggestions When RCM Is Not Applicable (Version 3)

In scenarios where user selections did not trigger RCM, a guidance section was added to educate users on correct parameter selection.

Key Features in V3:

1. A comparison table was introduced to show:
2. Required vs. user-selected parameters
3. ✓ / ✗ indicators for each condition
4. “How to Make RCM Applicable” section was added
5. Tool now helps users understand why RCM doesn’t apply and how to rectify their input to align with RCM triggers

Step 8: Enhancement – Legal Notification References (Final Version)

Upon stakeholder feedback, the tool was further refined to integrate notification numbers across all outputs, ensuring users have direct access to compliance references.

Final Additions:

1. Notification numbers appear:
2. In the initial RCM applicability result
3. Within the comparison and corrective sections
4. In the footer with the source citation: "Data Source: GST Notifications"

Final Tool Features

The AI-powered RCM Applicability Checker Tool now offers:

✅ Dropdown-driven selection of service, supplier, and recipient types

✅ Date-based logic for checking legal applicability

✅ Real-time result with legal references and notification details

✅ Explanatory section with corrective suggestions if RCM doesn’t apply

✅ Visual and tabular insights for easy interpretation

✅ Fully compliant with GST laws for FY 2024-25

🔗 Access the Tool Here:

https://claude.site/artifacts/9ee766bd-186b-42ba-9074-53c24c35cd24

Conclusion

This tool stands as a case study in leveraging AI for indirect tax compliance, demonstrating how prompt engineering and progressive iterations using generative AI can produce a user-centric and law-aligned utility. ICAI encourages members and students to explore and innovate using AI tools to support evolving compliance needs.