Your first import, start to finish: statement to sorted transactions

Almost everything Moneytaw does downstream — budgets, net worth, insights, your giving obligation — is built on one thing: transactions that are actually in the app and actually categorised correctly.

This is that flow, end to end. It takes about two minutes the first time and under thirty seconds after that.

Before you start: get the file

You need a statement from your bank. You almost certainly already have one — most Nigerian banks email a monthly statement as a PDF attachment. Search your inbox for "statement" and your bank's name.

Two things worth knowing before you go looking:

• Use the original file. The PDF or CSV your bank generated, not a screenshot and not a scan. Moneytaw reads the text inside the document; a picture of a statement has no text to read.
• Password-protected is fine. Most banks encrypt the attachment. Moneytaw will ask for the password and open it.

Supported today: GTBank, Access Bank, Zenith, UBA, First Bank, Sterling, Union Bank, Providus, Opay, Palmpay and Moniepoint. PDF works for all of them; Sterling also offers CSV and Opay offers Excel.

Step 1 — Upload it

Open Imports Manager, choose Statement Uploads, then Select a statement file. Pick the file. On desktop you can drag it straight onto the window.

What happens next is automatic:

1. Moneytaw detects which bank the statement came from, by reading its header.
2. It reads every transaction line — date, amount, narration, running balance.
3. It creates the account for you. You don't add the account first. The import makes it.

If auto-detection fails, you'll be asked to pick the bank yourself. That's the fallback, not the norm.

Step 2 — Review what came in

Your transactions are already categorised at this point. Four things ran, in order: your own rules first, then 166 built-in patterns written for Nigerian bank narrations, then AI for whatever those missed, and finally a fallback.

That last one matters. When Moneytaw isn't confident, it doesn't guess — it flags the transaction Needs review and leaves the category open rather than filing it somewhere plausible and wrong.

So scan the import for those flags. On a typical first statement you'll have a handful, usually transfers to individuals — because "TRF TO ADEBAYO O" genuinely could be rent, a loan, or a gift, and no amount of cleverness can tell which.

Step 3 — Fix one, and teach the rule

Tap a transaction that's wrong or unreviewed and set the right category.

Then Moneytaw asks: Teach a rule?

Automatically categorize future transactions as "Rent & Housing" when they match this keyword. The app guessed one below — edit it to match exactly what you want.

This dialog is the whole point of the article, so it's worth slowing down:

• Keyword — Moneytaw pre-fills a guess pulled from the narration, with the banking noise stripped out (channel codes, charge tokens, and similar are never useful as a match). Edit it. The pre-filled guess is a starting point, not a recommendation.
• Where to look — choose Narration, Sender or Receiver. Narration is the right answer most of the time.
• Create rule saves it. Not now applies your category to just this one transaction.

Pick the keyword carefully, because this is the decision that determines whether the rule ages well:

Too broad	Too narrow	About right
TRF	TRF TO ADEBAYO O 08/03 REF 4471	ADEBAYO O

A keyword that matches everything will mis-file things later. A keyword containing a reference number or a date will never match again.

Step 4 — Watch it apply itself

Saving the rule does two things.

It applies across the import you just did. Every other transaction matching that keyword recategorises on its own — you fix one, and the rest follow.

It runs on every future import. Rules are tier one in the categorisation waterfall, ahead of the built-in patterns and ahead of the AI. Nothing overrides them.

One deliberate limit: a rule never overwrites a transaction you categorised by hand. Your manual decisions are final. Automation cleans up after itself, not after you.

You can edit or delete rules anytime from Classification Rules in Settings.

Step 5 — See where everything landed

Two screens to check.

Accounts. The account Moneytaw created appears under Linked Accounts — the group for accounts created by an import, as opposed to Banks, which is where accounts you added by hand live. Its balance is calculated from the transactions themselves, not typed in, so it can't drift out of date.

Transactions. Everything from the import, grouped by date, with the month switcher at the top. You can filter by account, by type, by category, by amount range, by date range, and by review status — so "show me everything still needing review" is one tap.

That's the loop. Statement in, account created, categories mostly right, one rule taught, everything else sorts itself.

When something goes wrong

Four things account for nearly every failed import:

My bank isn't listed. Try a CSV export from your bank, or add the account manually and enter transactions. New banks get added regularly.

A statement won't import. Use the original PDF or CSV from your bank — not a screenshot, not a scanned copy — and check the bank is supported.

Some transactions are uncategorised. Open them from your history, set the right category, and save a rule so similar ones categorise themselves next time. This is working as designed, not a failure.

I see a duplicate. Delete the dataset from Statement Uploads and re-import it. If duplicates persist, contact support.

Do this once per account

Upload one statement from each bank you use. Each one creates its own account, and the balances add up into a single net worth.

The first import is the only one with any work in it. After that, the rules you taught are doing the sorting.