The Data Extractor is the fastest way to get detailed financial data into Kerdora. Upload a document, and the AI reads it and pulls out accounts, balances, income, insurance policies, and anything else it can find.
How I think about it
Clients don't always have their financial data organized in a spreadsheet. They have statements, tax returns, insurance policies, and pay stubs. The Data Extractor meets them where they are: hand us the documents, and we'll pull the data out.
It's especially useful for insurance policies and estate documents, the kinds of data that aggregation can't touch. A client can upload their life insurance policy and Kerdora extracts the carrier, benefit amount, premium, beneficiaries, and policy type without anyone typing a thing.
What it handles
The extractor works with almost any financial document:
Banking & investments — statements from banks, brokerages, retirement accounts
Insurance — life, disability, long-term care, property, auto, umbrella policies
Tax & income — tax returns, W-2s, 1099s, pay stubs, Social Security statements
Estate — wills, trusts, powers of attorney, health care proxies
Real estate — property documents, appraisals
Other — meeting notes, financial plans, retirement plan summaries
Supported file types: PDF, images (JPEG, PNG), Word documents, CSV, and plain text.
How it works
Upload — select one or more documents and submit. You can add context notes if the document needs explanation ("this is a joint account" or "this is from 2024").
Extraction — Kerdora's AI processes the documents in parallel. It identifies the document type, extracts a summary, and pulls out structured data (accounts, income, insurance, entities).
Review — you see everything the AI found, grouped by type. Each extracted item shows what it is, where it came from, and why the AI created it. You approve or reject each one individually.
Confirm — approved items become real data in the plan. Rejected items are discarded.
If the extraction gets something wrong, you can add clarifying context and re-run it. The AI uses your corrections to do a better job the second time.
Why the review step matters
The AI is good, but it's not perfect. A bank statement might list a transfer that looks like a separate account. A tax return might show last year's income, not current. The review step keeps you in control so nothing flows into the plan without your sign-off.
When you upload multiple documents at once, the extractor is smart enough to deduplicate. If three documents all reference "John Smith," it creates one entity, not three. Same with accounts that appear on multiple statements.