NineAnts · Categorisation
Bank transaction categorisation for treasury
Bank transaction categorisation assigns every bank transaction to a cash flow category, such as customer receipts, salaries or taxes. NineAnts categorises transactions with readable rules: the AI learns from your history and proposes rules, and your team approves them. In one implementation this cut manual categorisation work by 67%.
- Open any rule and see why
- Readable rules
- Your team approves
- AI proposes
- Is the training material
- Your history
- Applied in the morning sync
- Every day
The problem
Raw bank lines are noise
A bank statement tells you an amount, a date and a description written by someone else. It does not tell you that the payment was a customer receipt, a tax instalment or an intercompany transfer. Until each transaction has a category, you cannot report on cash flows, and you cannot forecast them.
Most treasury teams handle this in Excel: long keyword formulas, a lot of manual review, and a file that only one or two people fully understand. It works, but it takes time every day, it breaks when a supplier changes its payment reference, and it rarely covers more than part of the transactions.
The tricky cases make it worse. An intercompany payment looks like a normal supplier payment. Refunds and reversals flip the direction. The same supplier appears under different names depending on the bank.
How NineAnts does it
How NineAnts categorises your transactions
Rules you can read, written with the help of AI, approved by the people who know your cash flows.
Categories that fit your business
Typical categories include customer receipts, supplier payments, salaries, taxes, capex, intercompany, bank fees, interest, rent, utilities, loan drawdowns and repayments, and dividends. You define the categories that match how your business actually operates.
Tagging intercompany flows in their own category means they appear as normal flows at entity level and net to zero at group level. If they do not, that is your signal to look closer.
Learn from the history you already have
Most treasury teams already have years of categorised transactions in their files. That history is the training material. During implementation, NineAnts reads it, learns how your team thinks about its cash flows and builds a rule set that follows the same logic.
The AI proposes, your team approves
NineAnts categorises transactions with rules. Each rule is readable: you can open it and see exactly why a transaction received its category. There is no black box between your bank data and your reports.
Writing hundreds of rules by hand is the part nobody enjoys. NineAnts learns your data and proposes new rules. Your team reviews and approves them. The rule set stays yours, in your team's logic, and people stay in charge of it.
Our AI Agent helps here because it can read a transaction description the way a person does. It understands that "Office Lease Q2" is rent, where a pure keyword rule would need a specific entry for it.
Applied automatically, every day
Once the rules are in place, categorisation runs as part of your automation flows. A typical setup syncs every bank account at 07:00, applies the categorisation rules and refreshes Power BI before the team logs in.
Transaction categorisation is included in the Professional and Enterprise plans.
What you get
Clean categories, without the daily grind
- Every bank transaction in a cash flow category, automatically
- Readable rules you can open and check
- New rules proposed by AI and approved by your team
- A rule set that follows your team's own logic, learned from your history
- Categories defined around your business, including intercompany
- Daily categorisation as part of your automation flows
- A clean foundation for reporting, cash visibility and forecasting
- Included in the Professional and Enterprise plans
Case study
From Excel formulas to 89% automatic categorisation
- categorised automatically, up from 58% with Excel formulas
- 89%
- agreement with the team's own historical choices
- 99.3%
- less manual work, down to roughly a third
- 67%
A small example from one of our clients. Before NineAnts, their treasury team categorised bank transactions in Excel: an advanced model full of keyword formulas covered part of the transactions, and the rest was done by hand. It worked, but it cost the team a lot of time every day.
Over the years, the team had categorised around 180,000 transactions this way. Our agents read those files, learned the team's logic and built a new rule set. We then graded that rule set against all 181,031 decisions the team had already made.
The team still checks and manages the rules. The repetitive daily work is largely gone.
One platform
How it connects
A forecast built on badly categorised actuals inherits every one of those errors. Categories are what the forecasting models learn from: patterns per customer group, per supplier, per tax calendar, per entity. That is why every NineAnts implementation starts here.
- Collects for it
Data automation
Collects the bank and ERP data that gets categorised, and delivers the results to Excel or Power BI.
Learn more - Feeds it
Cash visibility
Collects the transactions from every bank and PSP.
Learn more - You are here
Transaction categorisation
A cash flow category on every transaction, from rules you can read.
- Learns from it
Cash flow forecasting
Uses the categories to forecast each cash flow type with its own model.
Learn more - Explains with it
Treasury AI Agent
Uses the categories to explain deviations and spot patterns.
Learn more
Built security first
Frequently asked questions
Straight answers to the questions treasurers ask us
What is bank transaction categorisation?
Bank transaction categorisation assigns each bank transaction to a cash flow category, such as customer receipts, supplier payments, salaries or taxes. It turns raw bank lines into information you can report and forecast on. In treasury, it is the step between collecting bank data and building a cash flow forecast. In American English it is usually called bank transaction categorization.
How does AI categorise bank transactions?
In NineAnts, AI learns from the transactions your team has already categorised and proposes rules that reproduce that logic. Our AI Agent can read transaction descriptions the way a person does, which helps with references that simple keyword rules miss. Your team reviews and approves each proposed rule, so every categorisation stays explainable.
Why use rules instead of letting AI categorise everything directly?
Rules are readable and predictable. You can open any rule and see why a transaction received its category, and the same transaction always gets the same result. That matters for audit and for trusting your forecast. In NineAnts, AI does the hard work of finding and proposing the rules, and your team keeps control over which rules are used.
How accurate is automated transaction categorisation?
It depends on your data and your categories. In one NineAnts implementation, a rule set learned from 181,031 historical transactions categorised 89% of transactions automatically, up from 58% with the client's Excel formulas. It agreed with the team's own historical choices 99.3% of the time, and manual categorisation work fell by 67%.
Can I categorise bank transactions in Excel?
Yes. Power Query can apply rule-based categorisation in Excel, and many treasury teams do this. It works, but it needs a lot of rules and a lot of maintenance, and coverage often stays partial. NineAnts can take over that logic, learn from your existing categorised history and apply the rules automatically every day, so the work no longer depends on one spreadsheet.
How do I handle intercompany transactions in categorisation?
Give intercompany flows their own category. At entity level they then appear as normal inflows and outflows, and at group level the intercompany category should net to zero. If it does not, you know where to look. NineAnts rules can pick up intercompany payments that would otherwise look like regular supplier payments.
Related reading
Blog
Aug 31, 2026 · 3 min read
67% less manual categorisation
During a recent NineAnts implementation, our categorisation engine learned a treasury team's own logic from 181,031 historical transactions and cut their manual categorisation work to roughly a third.
Read the articleBlog
Oct 10, 2024 · 5 min read
Excel for the Modern Corporate Treasury
Excel has been a trusted tool for treasury and finance professionals for decades. Most of the treasurers we speak to today still indicate that Excel is a vital tool in their treasury operations. It is widely used for key processes…
Read the articlePodcast
Episode 298 · Tracey Knight · 76 min
How Treasury Teams Automate Too Late and Rely on Spreadsheets Despite AI Growth
Tracey Knight has spent thirty years in and around corporate treasury. She explains when a company should first invest in treasury technology, and how to run a vendor demo that shows you what you actually need to see.
Listen to the episode
Find out how much of your categorisation can run automatically
Book a 30-minute session and bring an idea of how you categorise today. We will tell you honestly what NineAnts can take over.