Law Firm Cap Tables with Convertible Notes: Use case and AI Confidentiality
A cap table shows who owns a start-up and what a new round does to them, and the version a law firm builds before a seed round also converts the convertible notes and carries notes to the founders on what they will accept. The arithmetic is fiddly, so staff paste the sheet into ChatGPT or Claude to check the conversion or fill the percentages. This page marks every cell in a cap table that rule 9 and privilege bear on, recomputes the share counts and percentages, explains what each item exposes, and shows how the same sheet is built on Ermos inside the firm.
The cap table with convertible notes
Built by the solicitor for the founders before a priced round, to model the note conversion and to prepare the shareholders agreement and the resolutions.
| A | B | C | D | E | F | G | H | |
|---|---|---|---|---|---|---|---|---|
| 1 | Holder | Class | Pre-round shares | Pre-round % | Note converts to | Series Seed shares | Post-round shares | Post-round % |
| 2 | Ines Albrecht (founder, CEO)1 | Ordinary | 4,000,000 | 50.00% | 4,000,000 | 38.62% | ||
| 3 | Rohan Sandhu (founder, CTO)1 | Ordinary | 3,000,000 | 37.50% | 3,000,000 | 28.96% | ||
| 4 | Priyanka Venn (angel, 2025)2 | Ordinary | 400,000 | 5.00% | 400,000 | 3.86% | ||
| 5 | Employee option pool (unissued) | Options | 600,000 | 7.50% | 600,000 | 5.79% | ||
| 6 | Hartley Family Trust (Note A)3 | Seed preference | 212,000 | 212,000 | 2.05% | |||
| 7 | Dr Alan Quist (Note B)3 | Seed preference | 145,600 | 145,600 | 1.41% | |||
| 8 | Tallowwood Ventures Pty Ltd (lead)5 | Seed preference | 1,500,000 | 1,500,000 | 14.48% | |||
| 9 | Murray Pell (angel)2 | Seed preference | 500,000 | 500,000 | 4.83% | |||
| 10 | Total | 8,000,000 | 100.00% | 357,600 | 2,000,000 | 10,357,600 | 100.00% | |
| 11 | Note terms | A: $150,000 from 1 Mar 2026. B: $105,000 from 1 Jun 2026. 8% simple, whole months, to 30 Nov 20264 | $159,000 + $109,200 = $268,2007 | |||||
| 12 | Conversion price | Lower of $0.80 (20% off the $1.00 round price) and $0.75 ($6,000,000 cap / 8,000,000 shares)6 | $0.75 | |||||
| 13 | Internal | Term sheet says the investor takes 20%; it takes 19.31% because the notes convert on top of the $8,000,000 pre-money. Rohan will not go below 28%.8 |
An invented spreadsheet. Holders, amounts and share counts are invented and the arithmetic has been recomputed. Hover or tap any highlighted item for the analysis; the full list is below.
Cap table builder
An Agent+ that builds and updates a cap table from the documents on file: the share register, the convertible note deeds and the lead investor's term sheet. It sets out holdings before and after conversion and notes anything that does not reconcile between documents.
What it does
- Reads the register, note deeds and term sheet
- Builds the cap table and conversion scenarios in a spreadsheet
- Flags figures that differ between the source documents
What stays with your people. The solicitor checks every figure against the source documents. A partner reviews before the founders see the model.
An Agent+ is built once and anyone in the practice can run it. It works from your own documents and runs on Ermos Dominion, the AI unit in your office, or on Ermos Edge, your own instance in an Australian data centre.
What leaves the firm
- 1The founders and their holdings
Two named founders with roles and share counts. A company’s ownership before an unannounced round is confidential to the client under rule 9, and individuals’ shareholdings are personal information. Rule 9 confidentiality guide.
Frontier Models (ChatGPT, Claude) Transmitted to the provider, generally offshore on personal plans Ermos Edge Stays on your instance in Australia; leaves only inside a permitted excerpt, per question, to the named provider Ermos Dominion Stays on the unit in your office Use case Cap table builder. Builds the cap table from the share register, note terms and term sheet, with conversion scenarios laid out for the solicitor to check. See the use case - 2The angel investors
Named investors, when they came in and what they hold. Information about individuals who invested on the footing that it stays private. Privacy Act guide.
Frontier Models (ChatGPT, Claude) Named investors’ holdings transmitted Ermos Edge Stays on your instance in Australia; leaves only inside a permitted excerpt, per question, to the named provider Ermos Dominion Stays on the unit in your office Use case Cap table builder. Builds the cap table from the share register, note terms and term sheet, with conversion scenarios laid out for the solicitor to check. See the use case - 3The noteholders
A family trust and a named doctor who lent the company money. Their identities, and the fact that the company needed bridging finance, are confidential to the client. Rule 9 confidentiality guide.
Frontier Models (ChatGPT, Claude) Named lenders and the company’s bridging finance transmitted Ermos Edge Stays on your instance in Australia; leaves only inside a permitted excerpt, per question, to the named provider Ermos Dominion Stays on the unit in your office Use case Cap table builder. Builds the cap table from the share register, note terms and term sheet, with conversion scenarios laid out for the solicitor to check. See the use case - 4The note terms
Principal, issue dates, interest rate and the basis for counting interest. Commercial terms from confidential note deeds, and the inputs to every conversion figure in the sheet. Rule 9 confidentiality guide.
Frontier Models (ChatGPT, Claude) The note terms from confidential deeds transmitted Ermos Edge Stays on your instance in Australia; leaves only inside a permitted excerpt, per question, to the named provider Ermos Dominion Stays on the unit in your office Use case Cap table builder. Builds the cap table from the share register, note terms and term sheet, with conversion scenarios laid out for the solicitor to check. See the use case - 5The lead investor and the round
A named investor, the amount and the share count for a round that has not been announced. Confidential until completion, and a reader could size the company’s valuation from it. Rule 9 confidentiality guide.
Frontier Models (ChatGPT, Claude) An unannounced round and its lead investor transmitted Ermos Edge Stays on your instance in Australia; leaves only inside a permitted excerpt, per question, to the named provider Ermos Dominion Stays on the unit in your office Use case Cap table builder. Builds the cap table from the share register, note terms and term sheet, with conversion scenarios laid out for the solicitor to check. See the use case - 6The conversion price: discount and cap
The lower of a 20% discount to the round price and a price derived from a $6,000,000 valuation cap, here the cap price. The cap is a negotiated term in the note deeds and the number most investors ask about. Rule 9 confidentiality guide.
Frontier Models (ChatGPT, Claude) The valuation cap and the conversion price transmitted; on consumer tiers it can train future models Ermos Edge The same arithmetic either way; a solicitor must check the price against the note deeds Ermos Dominion The same arithmetic either way; a solicitor must check the price against the note deeds Use case Cap table builder. Builds the cap table from the share register, note terms and term sheet, with conversion scenarios laid out for the solicitor to check. See the use case - 7The converting amount
Principal plus accrued interest for each note. A derived figure, but the sheet shows its inputs. Rule 9 confidentiality guide.
Frontier Models (ChatGPT, Claude) The converting amounts and their inputs transmitted Ermos Edge The AI can mis-add interest or round shares wrongly; a solicitor must recompute Ermos Dominion The AI can mis-add interest or round shares wrongly; a solicitor must recompute Use case Cap table builder. Builds the cap table from the share register, note terms and term sheet, with conversion scenarios laid out for the solicitor to check. See the use case - 8The internal note
The investor’s stated 20% against the 19.31% the sheet produces, and the lowest stake a founder will accept. Privileged advice and the client’s negotiating floor in a cell labelled internal. Legal privilege guide.
Frontier Models (ChatGPT, Claude) The founder’s floor now on a third party’s servers Ermos Edge Stays on your instance in Australia; leaves only inside a permitted excerpt, per question, to the named provider Ermos Dominion Stays on the unit in your office Use case Cap table builder. Builds the cap table from the share register, note terms and term sheet, with conversion scenarios laid out for the solicitor to check. See the use case - 9The percentage formula
A formula that divides a holder’s post-round shares by the total in the total row. It holds no name or figure.
Frontier Models (ChatGPT, Claude) A formula with no name or figure in it; nothing to expose Ermos Edge The same formula either way; a person must check the range covers every holder Ermos Dominion The same formula either way; a person must check the range covers every holder Use case Cap table builder. Builds the cap table from the share register, note terms and term sheet, with conversion scenarios laid out for the solicitor to check. See the use case - 10A tab labelled confidential
A spreadsheet file can carry more than one tab, hidden columns, cell comments and author details. A paste of the visible cells leaves those behind, but uploading the workbook sends every tab, including the term sheet. Legal privilege guide.
Frontier Models (ChatGPT, Claude) An uploaded workbook carries every tab, including the confidential one Ermos Edge Stays on your instance in Australia; leaves only inside a permitted excerpt, per question, to the named provider Ermos Dominion Stays on the unit in your office Use case Cap table builder. Builds the cap table from the share register, note terms and term sheet, with conversion scenarios laid out for the solicitor to check. See the use case
What happens in Frontier Models (ChatGPT, Claude)
- The whole document is transmitted to the provider’s servers. On personal plans the user cannot choose where it is stored or processed, and for the major providers that has generally meant outside Australia, under that country’s legal process. Some business plans offer storage in Australia, but the provider still receives the content.
- On consumer tiers, prompts can be retained and used to train future models.
- The provider’s staff and systems can reach the content under the provider’s own policies.
- Once sent, the firm no longer controls it. Retention and deletion happen on the provider’s terms.
The same document on Ermos
Ermos Dominion
The document is drafted on a unit in your office, from your own templates and the matter file. Every marked item stays on the unit, which keeps working with the internet switched off. Nothing in it is transmitted to an AI provider, so there is no disclosure to analyse: the confidentiality work is the same as for your file server.
Ermos Edge
Your documents, the index and your access boundaries stay on your own instance in an Australian data centre. When a staff member asks for this draft, only their question and the excerpts they are permitted to see go to a named Australian provider, who is contractually barred from keeping the content or training any model on it. Nothing is processed outside Australia.
| Frontier Models (ChatGPT, Claude) | Ermos Edge | Ermos Dominion | |
|---|---|---|---|
| Where the document is processed | Provider servers, generally offshore on personal plans | Your instance in an Australian data centre; answers generated in Australia | The unit in your office |
| Who can read it | Provider staff and systems, under policy | Your staff, within the boundaries you set. The named Australian provider processes only permitted excerpts, under contract, and keeps nothing | Your staff only |
| Is it kept | On the provider’s terms | Kept on your instance; the provider is barred from retaining it | Stays on the unit |
| Can it train a model | On consumer tiers, yes | No, by contract | Only your own unit learns from it, on your hardware |
Part of Use cases in law firms.
Common questions
Do ChatGPT and Claude use the information in a cap table?
Is it safe to paste a cap table into ChatGPT or Claude?
What travels if we upload a cap table spreadsheet to ChatGPT or Claude?
How can a law firm use AI for a cap table without sending it to ChatGPT or Claude?
How does a convertible note convert into shares?
Why do cap table totals so often come out wrong?
What is the difference between a SAFE and a convertible note?
See Dominion answering from your own documents.
A demonstration takes 30 minutes, uses no client data, and comes with no obligation.
Book a demo