What if my export has columns Worthune does not use?

They are ignored. The mapping names only the columns the household record needs, and the draft mapping lists the columns it left unmapped so nothing is dropped by accident.

Can we import in stages?

Yes. Import one advisor's clients first, review them in the console and the client picture, then bring in the rest. Because re-runs skip what is already there, staging is safe.

Can AI make mistakes in the mapping?

The draft can be wrong, which is why it is only a draft: it passes a dry run over your sample rows before you see it, you review it, and deterministic code executes it. The AI never writes a household.

Help center

Bringing your book in

You bring a book in by exporting it from the system you use today, mapping its columns to the household record once, running a dry run that reports every problem row and field, and then importing for real. Imports can be re-run safely: a household already imported is recognized and skipped, never duplicated.

Last reviewed September 7, 2026. For the exact calls behind each step, see the developer docs.

Before you start

Decide what a household is in your world. Usually it is the client or family your CRM already groups, and that grouping's own identifier is the one to keep. Worthune stores it as the import key, and it is how a later re-run knows which rows it has already seen.

Then export. Most portfolio systems and CRMs export a spreadsheet or CSV with one row per account or per client. Either shape works; the mapping step tells Worthune which columns hold the balance, the account type, the owner's birth year and so on.

The steps

01Export the book from your current system to a CSV or spreadsheet.
02Map its columns to the household record. You can write the mapping by hand or paste ten sample rows and let Worthune draft it; the draft is checked before it is returned and you review it either way.
03Run a dry run. It writes nothing and returns a complete report, row by row and field by field, of what would be rejected and why. A book imports in batches of up to 200 households.
04Fix the rows the report names in your export or your mapping, and dry-run again until the report is clean.
05Run the import for real. Every row is validated; valid rows are imported even if others fail, and the response is the complete record of what happened.
06Re-run whenever you like. Rows already imported are skipped and report the household they became; nothing is overwritten by an import.

What a good import looks like

Every household carries the identifier your CRM uses, an advisor reference so the book rolls up by advisor from day one, and a label your team recognizes. Balances are by account and tax type. Missing information is allowed; the client picture will tell you what is missing and what filling it in would unlock.

What an import does not do

It never updates a household that already exists; updates go through the household's own record so they are versioned. It does not send notifications to your systems, so a large import does not flood them. And it does not hold positions or transactions, only balances by account.

Questions

Common questions

What if my export has columns Worthune does not use?: They are ignored. The mapping names only the columns the household record needs, and the draft mapping lists the columns it left unmapped so nothing is dropped by accident.
Can we import in stages?: Yes. Import one advisor's clients first, review them in the console and the client picture, then bring in the rest. Because re-runs skip what is already there, staging is safe.
Can AI make mistakes in the mapping?: The draft can be wrong, which is why it is only a draft: it passes a dry run over your sample rows before you see it, you review it, and deterministic code executes it. The AI never writes a household.

Something not covered here? Write to support@worthune.com.