Explore the build library
consultance.ai
Deal Copilot
Build libraryControlsAboutCase studiesBlogBook an audit
consultance.ai

Systems

WorkflowsDeal CopilotPrivate knowledgeFinance automationBuild library

Company

About usCase studiesBlogContact

Start a conversation

Free AI auditfounder@consultance.ai

Teams

Investment bankingCFOs & financePrivate equityPrivate creditReal estateProperty managementProfessional servicesOther industriesAll industries

Legal

Privacy policyTerms of service

Our approach

Audit the workflow.Build in your environment.Hand over a system you own.
Build something useful © 2026 consultance.ai
Build library/Family Office K1 Tie Out

Finance and data

Family Office K1 Tie Out

For family office controllers and CFOs: Claude reads each K1 and capital account statement, ties it to its entity in your book, and returns a cited exceptions list. Not tax advice.

Medium setup4 steps9 prompts
Open the setup guide Discuss an implementation
Finance and data

Start with the guided setup

Follow the setup below in your own workspace. For a demonstration of the prompt-copying process, open the optional walkthrough.

Watch the setup walkthrough
Step 1 · setup
Three minutes, four steps, nothing to install by hand

Claude sets it up for you. You just paste.

Never used Claude? It is free and takes 30 seconds to open. Copy the instruction below, paste it into Claude, and it reads this page and walks you through everything, one question at a time.

  1. Tell Claude how to talk to you

    One tap. It changes how much Claude explains, and how slowly it goes. You can change it any time.

  2. Copy your setup instruction

    A short instruction plus a link to this page lands on your clipboard. First copy asks for your email once. That unlocks every button across the whole library.

  3. Open Claude in a new tab

    Free account, no card, 30 seconds. This tab stays open so you can come back.

    Open claude.ai
  4. Paste, send, and answer one question

    Claude reads this page, asks one question about your work, then guides you step by step until your first output is right. If anything looks wrong, tell Claude what you see, and it fixes it with you.

▸Prefer the full prompt instead of the link? (optional)
Click to copy
I am comfortable copy-pasting and following instructions, but I am not a developer.
There is nothing to install for this one and no commands to type: it all happens inside Claude. If any instruction below implies a Terminal, translate it into the equivalent click path for me instead.
- Plain English. Define jargon the first time it appears.
- One step at a time, then wait for me to confirm before the next one.
- Tell me what success looks like at each step, and diagnose any error before moving on.

Follow the instructions below with those rules applied.

If you can browse the web, open and read this page in full first, it has the complete guide and every prompt you will run (the vault is under the-vault anchor): https://consultance.ai/library/family-office-k1-tie-out#the-vault . If you cannot open links, tell me and I will paste the page in, do not guess the prompts.

I want to set up the Family Office K1 Tie Out so Claude helps me read the K1s and capital account statements our funds send, tie each one to its entity and position in our consolidated book, and produce a cited exceptions list. It runs on my own files, in my own Claude. Walk me through it one step at a time and wait for me to confirm each step. Treat me like a controller or CFO who may never have built a Claude Project or used a terminal, and define every term once.

For a handful of funds this is NOT a Terminal or coding install. It is a private Claude Project with my files added. If I have a folder of many PDFs, or I want to repeat the run each time a K1 arrives, I should use Claude Code, and you will tell me.

Set expectations once, up front. The desk reads the files I give it and drafts. It does not connect to any fund portal, does not file anything and does not change my book. It is not tax advice: my CPA owns the returns, and a named person signs off every figure before it goes to the principal. Anything it cannot trace to a document is marked OPEN and becomes a question for the fund, the administrator or my CPA. If I want it wired into a standing process, the reply is "wire it" and consultance.ai sends a 30 minute slot.

Privacy, said plainly: the files go only to my own Claude account, never to consultance.ai. K1s carry tax identification numbers, so before I load them I should be on a Team or Enterprise plan, or have model training turned off in my personal plan (Settings, then Privacy, then under "Help Improve our AI models" turn the toggle off).

Ask me ONE question first and wait: where do the K1s, the capital account statements and the book extract live right now? (A) a few files I can upload (B) a folder of PDFs and exports on my machine, or files in a document store.

Path A, a handful of files: a private Claude Project.
1. In claude.ai choose Claude Opus 5.5 in the model picker. A Project is a private workspace with its own files that other chats cannot see.
2. Create a Project named "K1 tie out" plus the tax year. Add the K1s, the capital account statements and the book extract to the Project's knowledge, its file area.
3. Paste prompt 01 from the page and answer its questions. The first one asks what I expect to find. The last one is the person who signs off.
4. Run prompt 02 once. It is three fake funds with known answers. If one figure does not match the key, stop and tell me which, before any real file is used.
5. Then run prompts 04 to 09 in order, or ask a single question such as "which K1 replaced a number we already booked?" and run only the units that bear on it.

Path B, a folder of PDFs: Claude Code on my machine. Tell me plainly if I am on Path B before I start uploading files one at a time. It is Anthropic's command line tool, it needs a paid Claude plan, and it reads the folder and writes its outputs beside the files. The install page is https://code.claude.com/docs/en/quickstart . Give me one command at a time, say what success looks like, and wait. Then the same drill: prompt 01, prompt 02, then 04.

First session drill: prompt 01, then prompt 02 against its answer key, then prompt 04 on my real files. Prompt 03, the tools step, is optional and only for connecting my own systems read only.
Step 2 · run it on your data

Step 1 set it up. These 9 prompts do the work.

the vault

The 9 prompts

Grab the whole pack as one file, or tap any prompt below to copy it on its own. Placeholders that look like {{THIS}} get swapped for your own numbers — and if you ran Step 1, Claude fills them in for you.

One .md file · all 9 prompts, numbered, in order · nothing left out.
Click to copy
FAMILY OFFICE K1 TIE OUT: SHARED RULES
Every unit in this pack follows these rules. They are written here once.

1. WHERE TO RUN IT
- A handful of K1s and statements: the Claude app, in a private Project with the files added. Chat is fine here.
- A folder of K1s, statements and the book extract: Claude Code, opened in that folder. The outputs land as files beside them.
- The same run every tax season, or every time a K1 arrives: Claude Code, so the register is a file you update.
Model: select Claude Opus 5.5 in the model picker. Use Claude Sonnet 5 only for the first bulk read of a large folder of scanned PDFs.
Privacy: your files go only to your own Claude account, never to us. K1s carry tax identification numbers, so use Team or Enterprise, or turn model training off, before you add them.

2. BEFORE THE FIRST RUN
First ask what the human expects to find, and what worries them most, and record the answer. Then ask, one at a time:
- the family office name and the tax year
- the consolidated book extract: one row per position with entity, position name, carrying value at year end, contributions and distributions for the year, and any K1 income already booked
- the expected K1 list: one row per position with the fund, the fund's EIN, the holder entity and the holder's EIN. If there is none, build it from the book and label it ASSUMED
- last year's tax basis ending capital by position, if the office keeps one
- the tie tolerance per figure (default 1,000, label ASSUMED) and the size of change worth a question (default 10,000, label ASSUMED)
- the named person who signs off
A missing item gets that plain default, labelled ASSUMED, except documents, which are never assumed.

3. STOP RATHER THAN GUESS
- A file path or folder listing pasted into chat: this is a Claude Code job. Say so and stop.
- A truncated, rotated or unreadable page: name it, quote the characters you can read, and do not fill the figure. Mark it UNREAD.
- A K1 or statement you do not have: ask once, name it, mark it OPEN, and say who holds it (the fund, the administrator, the CPA).
A stop blocks only the outputs that need the missing input. Every other check still runs.

4. LABELS ON EVERY FIGURE
REPORTED (document, page) · CALCULATED (show the arithmetic) · ASSUMED (who set it) · MARKETING (a claim in a letter or cover note: it makes a question, never a number) · NOT IN SOURCE.
An estimate letter or an estimated K1 is a document, not a K1. Its numbers are REPORTED as estimates and never become the tax capital figure.

5. THE TIES, IN ONE PLACE
T1 Item L foots: beginning + contributions + net income (loss) minus distributions = ending, within tolerance.
T2 Item L beginning = last year's tax basis ending capital.
T3 K1 Item L figures = the tax basis column of the fund's capital account statement for the same version of the K1.
T4 The GAAP or NAV column ending capital of the statement = the book carrying value. Its contributions and distributions = the book's for the year.
T5 The gap between the GAAP and tax columns is explained by the statement's own bridge, and the bridge adds up.
T6 The holder name and EIN on the K1 = the holder entity the book and the expected list assign to that position.
T1 to T3 and T6 are arithmetic or identity. T4 and T5 are the bridge between two bases. Tax basis capital is not expected to equal the book value. It is expected to be explained.

6. THREE FLAG TAGS
- CHANGES THE DECISION: a figure or an entity assignment that would change what the principal is told. Only these reach the verdict.
- WORTH A QUESTION: only if the answer could move a figure by more than the question threshold. Write the question and who answers it.
- EXPLAINED BY CONTEXT: one line under "Checked, normal". Once closed, it stays closed and never returns.
Test every flag against the list below before you raise it.
HARD BLOCK is separate: only a number that does not tie (T1, T3 or T4 outside tolerance with no disclosed reason) or an identity that does not match (T6). A block stops the verdict and the total for that position, never the analysis. Keep running every other check, and always print what happens by default if the human does nothing. A missing document is a question, never a block.

7. LOOKS WRONG, IS NORMAL
- Tax capital differs from GAAP or NAV capital, and from the book carrying value | different bases by design | not normal when the statement gives no bridge or the bridge does not add up
- Negative Box 1, Box 2 or other loss lines | funds report losses | not normal when Item L does not foot
- Ending capital of zero on a K1 marked final | the position was fully redeemed | not normal when the book still carries a value
- A K1 not yet issued, with an estimate letter | fund of funds and real asset funds report late | the estimate is OPEN, never a figure
- A corrected K1 supersedes an earlier one | the later issue date wins, the earlier one is history | it still changes any figure already booked, so compare it to the booked figure
- GAAP distributions higher than K1 distributions | declared in December, paid in January, disclosed on the statement | not normal when no note explains the gap
- The holder spelled three ways on K1, statement and book | match on EIN first, name second | not normal when the EIN points to a different entity
- Capital called and distributed the same day | the book may show net, the K1 shows gross on both lines | not normal when the two sides do not match each other
- Item L beginning differs from last year's ending | last year's K1 was later corrected | not normal when no corrected K1 exists
- Item L labelled GAAP, 704(b) or other instead of tax basis | funds choose the method | ask which basis, do not compare it to a tax basis statement as if it were the same

8. STANCE
A neutral colleague who has done this many times. Do not hunt for problems. Do not take a cover note as proof. Do not explain away arithmetic that does not tie.

9. HANDOFF (every unit ends with this)
Read (and not read) · Outputs · Checks T1 to T6: pass, fail or not run · Open items, each with the document needed and who holds it · Next unit · Sign off needed from the named person.
Click to copy
Run prompts 04 to 09 on this sample before you use real data. Compare your result to the answer key. If one figure does not match, stop and find out why first. One K1 is missing on purpose, so you see the desk stop on it instead of guessing.

Family office: Reed Family Office. Tax year 2025. Tolerance 1,000. Question threshold 10,000. Sign off: the named controller. The view going in: all three funds have reported.

BOOK (consolidated book, page 1)
A  Reed Family Trust       Alder Peak Partners LP        carrying 1,120,000  contributions 0       distributions 80,000   K1 income booked 120,000
B  Reed Family Trust       Birch Hollow Credit Fund LP   carrying 619,000    contributions 50,000  distributions 20,000   K1 income booked 38,000
C  Reed Family Trust       Cobalt Ridge Fund LP          carrying 312,000    contributions 0       distributions 0        K1 income booked none
EXPECTED K1 LIST (page 2): A, B and C, holder Reed Family Trust, holder EIN 551000001.
LAST YEAR TAX BASIS ENDING CAPITAL (page 3): A 900,000  B 500,000  C 300,000

ALDER PEAK K1 (issued 3/10/2026, pages 1 and 2). Holder Reed Family Trust EIN 551000001. Item L tax basis: beginning 900,000, contributed 0, net income 120,000, distributions 80,000, ending 940,000. Box 1 (30,000). Box 9a 150,000.
ALDER PEAK STATEMENT (page 1) investor "Reed Fam Trust". GAAP: beginning 1,050,000, contributions 0, income 150,000, distributions 80,000, ending 1,120,000. Tax: beginning 900,000, contributions 0, income 120,000, distributions 80,000, ending 940,000. Page 2: difference 180,000, unrealised gains not taxed until sale.
BIRCH HOLLOW K1 (issued 3/12/2026, page 2). Holder Reed Family Trust EIN 551000001. Item L: beginning 500,000, contributed 50,000, net income 38,000, distributions 20,000, ending 568,000. Box 5 38,000.
BIRCH HOLLOW CORRECTED K1 (issued 4/20/2026, page 2, marked "supersedes the K1 issued 3/12/2026"). Item L: beginning 500,000, contributed 50,000, net income 63,000, distributions 20,000, ending 593,000. Box 5 63,000.
BIRCH HOLLOW STATEMENT (issued 4/25/2026, page 1). GAAP: beginning 540,000, contributions 50,000, income 49,000, distributions 20,000, ending 619,000. Tax: beginning 500,000, contributions 50,000, income 63,000, distributions 20,000, ending 593,000. Page 2: difference 26,000, accrued income timing.
COBALT RIDGE STATEMENT (page 1) only. GAAP: beginning 300,000, contributions 0, income 12,000, distributions 0, ending 312,000. No K1 and no estimate letter has arrived.

ANSWER KEY
Alder Peak: T1 940,000 = 900,000 + 120,000 minus 80,000, pass. T2 pass. T3 pass. T4 1,120,000 = book, pass. T5 180,000 = 1,120,000 minus 940,000, bridged, pass. T6 pass on EIN. Normal, EXPLAINED BY CONTEXT: the loss in Box 1 (T1 still foots), the tax versus GAAP gap, and "Reed Fam Trust" against "Reed Family Trust" (same EIN). No flag.
Birch Hollow: the corrected K1 of 4/20/2026 governs. Net income 63,000, ending 593,000, T1 pass (500,000 + 50,000 + 63,000 minus 20,000), ties to the statement tax column. The book holds 38,000. Restate by plus 25,000, above the 10,000 threshold. Tag CHANGES THE DECISION. The trap is using the first K1 (568,000). T4 619,000 = book, pass.
Cobalt Ridge: OPEN. No K1, no estimate letter. Ask the fund and name who holds it. T3 not run. T4 312,000 = book, pass. No tax figure is invented. A missing document is a question, not a block.
Verdict: HOLD, because of the Birch Hollow restatement and not because of Cobalt Ridge. Default if nobody acts: the principal note carries a status table and no total.
Tied tax capital so far: 940,000 + 593,000 = 1,533,000, shown as "excludes Cobalt Ridge". Book carrying total 2,051,000.
A run that flags a normal item, uses the first Birch Hollow K1, or gives Cobalt Ridge a tax capital figure, has failed.
Click to copy
<unit name="tools-setup">
Follow the shared rules in prompt 01.
What it does: suggests tools for the jobs in this pack, then installs only the ones the human approves, on the human's own accounts.
Use when: after prompt 02 passes, before the first real run. Skip it if upload or paste is enough.
Needs: the "Needs" lines of every unit in this pack, and the systems the human already uses (your fund administrator portals, your document store, and your accounting or portfolio system).
Method:
1. See what is already connected. In Claude Code, run `claude mcp list`. In the Claude app, look up the current steps in Anthropic's own help pages with your web search, then ask the human what their Claude already shows as connected.
2. For each need, search the web yourself for an MCP server or connector published by the vendor the human already uses, or listed by Anthropic. Prefer: official publisher, read-only access, sign-in by OAuth, updated in the last 6 months.
3. Show one table: need · tool · publisher · official page URL · read only or read and write · how far it reaches (one folder, one account, or everything the human can open) · what it costs on the human's own plan (write "not stated" if the page does not say) · last update (write "no date shown" if none).
4. For each tool the human approves, open its official page and follow its own steps. There are three kinds:
   minus A server with a command or URL on its page: copy the command word for word. Never write one from memory. In Claude Code the shape is `claude mcp add --transport http <name> <url>` for a remote server, or `claude mcp add <name> -- <command>` for a local one. Use `--scope project` when the tool is for this folder only.
   minus A connector added in the Claude app (no command): give the human the steps from the official help page, as written there. Claude Code can then use connectors added to their claude.ai account.
   minus An official page with no command, URL or steps, or steps behind a login: mark it ASK VENDOR, and write the one question to send their vendor contact or admin. Use UPLOAD for that need until it is answered.
5. Install one at a time. Then make one small read-only test call and show the result. Then go to the next tool.
Rules: use only the human's accounts, keys and credits. Never ask for a key in chat. Use the tool's own sign-in, or an environment variable the human sets. Start every tool read-only. If the tool has no read-only setting, connect it with a user or role that can only view, or tell the human it can write and let them decide. If it reaches further than this job needs, say so before they approve it.
<trap>A server that is not from the vendor, asks for write access or broad scopes, or sends documents to a third party for "processing". Leave it out, even when it is free.</trap>
Stop: no official page means do not install it. Mark that need UPLOAD and carry on. A need whose data lives in a document, not a system (an agreement, a memo), is UPLOAD from the start. Every other unit still runs.
Output: the table, what was installed, each test result, and the handoff. Next: the first working unit.
</unit>
~~~
~~~
claude mcp add --transport http sentry https://mcp.sentry.dev/mcp
claude mcp add my-server -e API_KEY=xxx -- npx my-mcp-server
-s, --scope <scope>  Configuration scope (local, user, or project) (default: "local")
Click to copy
<unit name="k1-register">
Follow the shared rules in prompt 01.
What it does: reads every K1, corrected K1, estimate letter and capital account statement, and writes one register row per document.
Use when: first, after prompt 01 is answered. Run it again when a new K1 arrives.
Needs: every K1, estimate letter and capital account statement, and the expected K1 list.
Method:
1. One row per document: fund, fund EIN, holder name and EIN as printed, document type (K1, corrected K1, estimate letter, statement), issue date, and the page every figure came from.
2. From each K1 take Item L (beginning, contributions, net income or loss, distributions, ending), the basis the box says it uses, and Boxes 1, 2, 5, 6a, 8, 9a, 11, 13, 18, 19A. Take only what is printed. Mark an unreadable figure UNREAD.
3. From each statement take both columns, GAAP or NAV and tax basis, the stated difference and its reason, and any note about timing.
4. Set the status of each K1: FINAL, CORRECTED (names the K1 it supersedes), ESTIMATE, or NOT RECEIVED (on the expected list, no document).
5. Check every row against the expected list: any fund on the list with no K1 is NOT RECEIVED, and any K1 not on the list is a question.
<trap>An estimate letter looks like a K1 and carries a figure. It is not a K1. Its number is REPORTED as an estimate and never enters the tax column.</trap>
Stop: an unreadable page blocks that figure only. A position with no document at all is NOT RECEIVED, and every other row still runs.
Output: the register table, with a page citation on every figure, then the handoff. Next: 05 entity match.
</unit>
Click to copy
<unit name="entity-match">
Follow the shared rules in prompt 01.
What it does: ties each K1 and statement to the holder entity and position in the consolidated book.
Use when: after 04.
Needs: the register from 04, the book extract, the expected K1 list.
Method:
1. Match the fund to a book position on EIN first, then on name. The same fund may be spelled three ways.
2. Match the holder on the K1, then on the statement, to the entity the book and the expected list assign. EIN first, name second. This is T6.
3. Where the K1 holder EIN is a different entity than the book, do not assign the K1 to either. List both candidates, the amounts at stake, and the question for the fund.
4. List every book position with no K1 and no estimate, and every K1 with no book position.
<trap>A name that looks right with an EIN that belongs to a sibling entity. Names are matched by people, EINs by the IRS. The EIN decides.</trap>
Stop: no expected list means build one from the book and label it ASSUMED. Say so, and carry on.
Output: the match table (document, entity, position, basis of the match), the identity problems, then the handoff. Next: 06 capital tie.
</unit>
Click to copy
<unit name="capital-tie">
Follow the shared rules in prompt 01.
What it does: runs ties T1 to T5 for every matched position and shows the difference to the book by entity.
Use when: after 05.
Needs: the register, the match table, the book extract, last year's tax basis ending capital.
Method:
1. T1: recompute Item L from its own lines and compare to the printed ending, with the arithmetic shown.
2. T2: beginning against last year's ending. A difference is explained only by a corrected prior year K1 you hold.
3. T3: each K1 Item L figure against the tax column of the statement issued for the same version. If the statement pre dates a corrected K1, say so and ask for the newer one.
4. T4: GAAP or NAV ending against the book carrying value, and the statement's contributions and distributions against the book's. A timing note on the statement can explain a distributions difference, and it must name the amount and the date.
5. T5: GAAP less tax ending capital equals the statement's stated difference, and the stated reason can be real. Do not judge the reason beyond that.
6. By entity: book carrying value, tax basis capital, the difference, and the status of every position.
<trap>Comparing K1 tax capital straight to the book carrying value and calling the gap an error. The gap is expected. The tie is to the statement's tax column, and the bridge explains the rest.</trap>
Stop: a position with no K1 runs T4 only. T1 to T3 are marked NOT RUN, with the document needed. Nothing else stops.
Output: one table per entity with every tie as PASS, FAIL or NOT RUN and the difference in dollars, then the handoff. Next: 07 corrected and late.
</unit>
Click to copy
<unit name="corrected-and-late">
Follow the shared rules in prompt 01.
What it does: finds every K1 that replaces a number already booked, and every K1 still to come.
Use when: after 06, and each time a K1 or a corrected K1 arrives.
Needs: the register, the book's booked K1 income by position, the expected K1 list.
Method:
1. For every position with more than one K1, the latest issue date governs. List the earlier K1 as superseded.
2. Compare the governing K1 to the figure already booked: net income and ending capital. Show the dollar change and the restatement needed.
3. Tag a change above the question threshold CHANGES THE DECISION. A change under the threshold is a line in the register and a WORTH A QUESTION only if the answer could move a figure above the threshold.
4. For every K1 not yet issued: the fund, the entity it flows to, the estimate letter if any, the date the fund gave, and who to ask. Estimated figures are listed as estimates and are not totalled.
<trap>Booking the first K1 because it is the one on the file. A corrected K1 is a replacement, not an addition, and the statement may pre date it.</trap>
Stop: a corrected K1 that does not name the K1 it replaces is a question to the fund. Everything else still runs.
Output: the superseded list, the restatement table, the late K1 list by fund and entity, then the handoff. Next: 08 exceptions.
</unit>
Click to copy
<unit name="exceptions-list">
Follow the shared rules in prompt 01.
What it does: turns every result from 04 to 07 into one cited exceptions list, each line tagged and tested against the normal list.
Use when: after 07.
Needs: the outputs of 04 to 07 and the normal list in prompt 01.
Method:
1. Test each candidate flag against the normal list in prompt 01 before it is raised. A flag that matches a normal row is EXPLAINED BY CONTEXT and goes on one "Checked, normal" line.
2. Tag the rest: CHANGES THE DECISION, WORTH A QUESTION, or HARD BLOCK. Every line cites the document and page for both sides of the number, and the arithmetic.
3. For each line write the one thing that closes it and who holds it: a corrected K1, a confirmation from the fund, a booking entry for the controller.
4. Give the verdict: READY, READY WITH OPEN ITEMS, or HOLD. HOLD follows only a CHANGES THE DECISION flag or a HARD BLOCK. An OPEN document alone gives READY WITH OPEN ITEMS.
5. Print what happens by default if the human does nothing.
<trap>Raising the tax versus GAAP gap, a loss line or a disclosed timing difference as a break. They are normal here, and each one you raise buries the one that is real.</trap>
Stop: none. A missing document is already an OPEN line, so this unit always runs.
Output: the exceptions table (tag, position, entity, what, the figure on each side, citation, what closes it, who), the "Checked, normal" line, the verdict and the default, then the handoff. Next: 09 recount and principal note.
</unit>
Click to copy
<unit name="recount-and-note">
Follow the shared rules in prompt 01.
What it does: recounts the key figures from the source documents before it reads the tables, then writes the note for the principal.
Use when: last. Nothing goes to the principal before this unit.
Needs: the source documents, and the outputs of 06 to 08.
Method:
1. Without looking at the tables, recompute from the source documents: Item L for every K1 that governs, the tax basis ending capital for every tied position, and the total.
2. Compare to the tables. If any figure differs by more than the tolerance, print RECOUNT FAILED, show both numbers, and stop here. Write no note.
3. Otherwise write the note for the principal in plain paragraphs: what the family owns at year end on the book basis, what tax basis capital is for the positions that tie, what changed since it was last told, and what is still open.
4. Every figure carries its source page. A total is shown only with the list of positions it excludes, by name.
5. If any HARD BLOCK or CHANGES THE DECISION flag is open, the note carries a status table and no total, and says so in the first line.
<trap>Writing the headline number from the tables because they were just produced. The recount reads the K1s again. A total that nobody re added is not a total.</trap>
Stop: RECOUNT FAILED blocks the note. Everything else already ran.
Output: the recount table, the note for the principal, the handoff, and the sign off line for the named person. Next: sign off by the named person, outside Claude.
</unit>

From a working build to an operating system.

For family office controllers and CFOs: Claude reads each K1 and capital account statement, ties it to its entity in your book, and returns a cited exceptions list

Path A · free

Run the build independently

The setup rail and every prompt above are free and stay free. The cost is your time, and the risk of wiring it wrong on live data.

Back to the prompts
Path B · done with you

Connect it to your operation

We would deliver a private K1 intake desk: the units loaded into your Claude account, your book extract and expected K1 list wired in read only, a register that updates when a K1 or a corrected K1 arrives, and a sign off step so no untied figure reaches the principal. Reply "wire it" for a 30 minute slot.

Book a build call
data safety

Before you use live numbers

  • • Run last quarter's numbers first. Live data is not a test bed.
  • • Nothing here uploads to us. It runs in your own Claude account, on your own machine.
  • • A named human reviews and signs every output before it reaches a board, lender, or client.
  • • Mask account numbers and names to the minimum the task needs.
the fine print

Straight answers on ownership

Pack authored by consultance.ai. Your K1s and statements stay in your own Claude account, we never see them. This reads, ties and flags. It is not tax, accounting or legal advice: your CPA owns the returns, and a named person signs off every figure before it reaches the principal.

Bring this workflow into your operation.

Book a build callBack to the library

Want this wired into your stack instead of running it yourself? That is our AI deal desk and finance automation service.

Newsletter

New builds, as they ship

One short note when a new build ships: what it does, who it is for, and how to run it in your own Claude.

Free. Three short welcome emails over your first three days, then at most one email a week. Unsubscribe in one click.

More like this

Related builds

All builds
Finance and data
Finance and data

Month End Close AI Agent Team

For controllers and CFOs: a 12 agent month end close team in your own Claude that preps reconciliations, accruals, entries and flux notes as a structured review. You approve every ledger post.

Setup guide
Finance and data
Finance and data

RIA Back Office AI Agent Team

For RIA principals, independent advisers and wealth firm COOs and CCOs: 10 ready made Claude Code agents that work as a screening checklist for drift, rebalancing, tax loss scans, trades against the client's file and Marketing Rule drafts. A named adviser approves anything that moves money.

Setup guide
Finance and data
Finance and data

Ramp Claude Connector Setup Guide

For controllers and finance ops teams on Ramp: Claude setup steps plus 12 read only prompts for spend baseline, close readiness, accruals and flux. Replaces the manual Ramp export to Excel.

Setup guide
Get the free kitBook a call

From independent build to integrated workflow

What you have

The free version above. It runs this read on a file you paste in, in your own Claude. Yours to keep.

What we install

The same read, wired into your cfo and back office stack, running the second a report lands. Nothing to upload, no new login.

See the installed version

Share with your workflow owner.

Bring the relevant team into the conversation with the build and setup guide.

Forward by email
in one line

What is Family Office K1 Tie Out?

Family Office K1 Tie Out is a finance and data build in the consultance.ai AI Build Library. For family office controllers and CFOs: Claude reads each K1 and capital account statement, ties it to its entity in your book, and returns a cited exceptions list. Not tax advice. It fits Controllers, CFOs and finance leads at single and multi family offices who receive K1s and capital account statements from many funds at different times, some corrected, and have to agree them to the consolidated book. Setup difficulty is Medium, with 4 plain-English steps.

What does Family Office K1 Tie Out do?

For family office controllers and CFOs: Claude reads each K1 and capital account statement, ties it to its entity in your book, and returns a cited exceptions list. Not tax advice.

Who is Family Office K1 Tie Out for?

It fits Controllers, CFOs and finance leads at single and multi family offices who receive K1s and capital account statements from many funds at different times, some corrected, and have to agree them to the consolidated book.

How hard is Family Office K1 Tie Out to set up?

Medium to set up — one guided setup instruction covering 4 plain-English steps, plus 9 ready-to-run prompts on the resource page.

How would consultance.ai build this out?

We would deliver a private K1 intake desk: the units loaded into your Claude account, your book extract and expected K1 list wired in read only, a register that updates when a K1 or a corrected K1 arrives, and a sign off step so no untied figure reaches the principal. Reply "wire it" for a 30 minute slot.

What are the licensing terms?

Pack authored by consultance.ai. Your K1s and statements stay in your own Claude account, we never see them. This reads, ties and flags. It is not tax, accounting or legal advice: your CPA owns the returns, and a named person signs off every figure before it reaches the principal.

Connect this build to your workflow.

Discuss how Family Office K1 Tie Out fits your systems, controls, and team. We define the integration, review process, and deployment requirements with you.

This build comes from our AI consulting and AI implementation practice — see the full AI in finance guide and how we work with CFO teams.

Book an audit