Learned Patterns and Smart Categorisation
How NivoaFlow learns from your manual categorisations and auto-applies them to future transactions.
Overview
NivoaFlow watches how you categorise transactions and gradually learns your habits. After you assign the same merchant to the same category enough times, NivoaFlow turns that into a learned pattern and starts applying the category automatically to future transactions from that merchant.
Available on Personal and Household tiers.
How Learning Works
When you manually categorise a transaction (or correct one we got wrong), NivoaFlow records the merchant and category pair. Once it has seen the same merchant categorised the same way three or more times consistently, the pattern flips to active and starts auto-applying.
You will see two states on the Learned Patterns page:
- Learning — we have noticed the pattern but are not auto-applying yet
- Active — we are confident enough to auto-apply this category to new transactions from this merchant
Each pattern shows a confidence percentage and the number of times we have observed it.
Where to View Your Patterns
- Go to Settings > Learned Patterns
- Toggle between All and Active Only to filter the list
- Each row shows the merchant, the category we learned, the status (Learning / Active), and our confidence
Deleting a Bad Pattern
If we have learned the wrong pattern — for example, we are auto-categorising a clothing store as Groceries because of a one-off mistake — remove it:
- Open Settings > Learned Patterns
- Find the offending pattern
- Select the trash icon on the right
- The pattern is deleted immediately and we stop auto-applying that category
The next time you correct a transaction from that merchant, learning starts fresh.
Learned Patterns vs. Categorisation Rules
NivoaFlow has two different ways to automate categorisation. They work side by side and serve different purposes.
| Learned Patterns | Categorisation Rules | |
|---|---|---|
| Created by | NivoaFlow, automatically | You, manually |
| Trigger | Repeated manual categorisations | A pattern you define (text, regex, amount, account) |
| Best for | Letting the app figure out routine spending | Forcing a specific outcome you always want |
| Override | Delete the pattern | Edit or delete the rule |
If a learned pattern and a rule conflict, rules win — they are explicit and you wrote them on purpose.
Why Was a Transaction Auto-Categorised?
Open the transaction detail view. If the category came from a learned pattern, you will see an indicator showing it was assigned automatically. You can change the category right there — your correction feeds back into the learning system, and if you correct the same merchant a few times, the pattern updates or gets retired.
Tips
- Review Learned Patterns every few weeks and clean up anything that looks off
- For categories you always want enforced (such as a specific salary deposit going to Income), create an explicit rule instead of relying on learning
- Learning works on the merchant name, so if a single merchant covers multiple kinds of spending (e.g., a big-box store), prefer rules with amount conditions
Related articles
Categorising Transactions
How NivoaFlow assigns categories using rules, learned patterns, and machine learning, with full manual control.
Creating Categorisation Rules
Set up pattern-matching rules to automatically categorise transactions based on merchant names and conditions.
Bulk Actions and Review
Use multi-select mode to batch categorise, review, and manage transactions efficiently.