Ai Tools for Accountants #4 – Client Structure Builder

If you ask ChatGPT to produce a diagram of a company structure with a Trading Company, shareholder trust (with a company trustee), this is what you get:

Diagram (Text/ASCII Representation)

       Barry           Pam

        |               |

      (50\%)           (50\%)

        \               /

          ---------------

                |

      Barry FT Co Pty Ltd (Trustee)

        |             |

    (Director)    (Director)

     Barry         Pam

                [Trustee]

                |

      Barry Family Trust

            |

     (100\% Shareholder)

            |

   Trading Co Pty Ltd

        |

    (Director)

      Barry

That’s disappointing, as the ability to produce clean, clear and consistent client business and investment structure visuals would be a great help for communication and education, both internally (staff trying to grasp general and specific structuring decisions and issues) and externally (clients having a clear image of their asset placements, profit flows and safe investment places).

Of course there is a way for Ai to produce clean, clear and consistent client business and investment structure visuals; here’s how we went about it.

A word on Privacy….

Our Client Structure Builder involves the disclosure of client business information, entity names, ownership, Directorships etc. We have opted-out of storage and data-training with our enterprise subscriptions to Ai models, meaning client information is not stored or used for training the model post analysis; the data (or in this case, the information relevant to the Client Structure Builder) is deleted with only the structure output remaining.

What design system to use?

Different people will have different views on this, but the design systems we used has the following rules:

1)         The business/investment structure has a top down-approach i.e. ownership at the top, subsidiaries cascading down the page

2)         Entities within the structure are annotated as follows:

Individual non-Director: Bold “X”

Individual Director: Bold “X” in a bold circle

Company: Rectangle with width 2x height

Discretionary Trust: Equilateral triangle with point at top

Fixed/Unit Trust: Equilateral triangle with point at bottom

Business: Oval, placed just below the business owner(s)

Partnership: Dotted Oval around partners to the partnership

3)         Ownership interests: Full line

4)         Debt Interests: Dotted line (optional)

Which model and language to use?

The major language models used on the Elfworks platform – Claude, ChatGPT, Gemini and Grok – cannot produce clean, consistent diagrams (at the moment). Thankfully, someone created the Mermaid programming language, a JavaScript based diagramming and charting tool that allows for consistent production of diagrams. We linked Mermaid into a detailed prompt for Gemini 2.5 for the cleanest and most consistent results.

How long does the Client Structure Builder take?

It depends on the structure. The Structure Builder requires information about an entity such as name, trading/investment, shareholding/ownership, Directorships (for companies) and type of trustee (for trusts). For a business structure with a trading company, holding trust, company trustee and company beneficiary, this will take around 10 minutes assuming you have all the required information ready to go. Note: We decided to build a separate family trust distribution tax (FTE and IEE) risk tool, separate from the Client Structure Builder, as the “family group” visuals alongside the business/investment structure were far too crowded.

If you would like full, free, no obligation trial of an Ai platform designed from the bottom-up for Australian accountants with a Client Structure Builder (and dozens of others), please contact us at info@elfworks.ai or phone me on 0418 902 440.

Thank you for reading. Next blog I’ll take a look at Ai Tools for Accountants #5 – Family Trust Distribution Tax Risk tool.

Please click here for Elfworks pricing information.