Data Agent

Transform Financial Data. Automatically.

Numra's Data Agent takes messy financial data from any source, applies your rules, and delivers clean, structured outputs ready for your ERP, reports, or team.

Data Agent — Transformation Jobs 2 ready for review
Last run 6 min ago
Emma Byrne €84.30
Uber Eats · 6200 — Meals & Entertainment
VAT on Purchases 23% Completed
Lisa Gregan €187.40
Ryanair · 6100 — Travel
Zero Rated Purchases Review
Conor Doyle €214.88
Amazon · 6300 — Office Supplies
VAT on Purchases 23% Completed
Tom Whelan €59.99
Adobe · 6400 — Software & IT
EU Reverse Charge 23% Completed
6 expenses · 2 ready for review View all transactions →

This Isn't an Off-the-Shelf Tool

What you see below is a representation of what's possible. Every Data Agent is built from scratch around the specific data flows your team needs.

What We Customise Fully Tailored
Data Connections
Bank feeds, ERPs, payment platforms, spreadsheets, PDFs — whatever your sources and destinations, we connect to them.
Transformation Rules
GL mapping, currency conversion, field formatting, entity splitting — your logic, in plain English.
Output Formats
CSV, Excel, ERP-ready imports, custom templates — outputs built to match exactly what your destination system expects.
Scheduling & Triggers
On-demand, on a schedule, or triggered automatically when new data arrives. You decide when jobs run.
Data Flows
How data moves from source to output, which fields map where, what gets split, merged, or recalculated along the way.
Controls & Exceptions
Validation checks, error thresholds, missing data flags — built around the problems that break your current process.

How the Data Agent Works

From raw data to clean output. Every transformation visible. Every decision explained.

Integrations
Rules
Jobs
Outputs
Logs
Revolut
Corporate Card Account
Connected
Stripe
Acme Corp — Production
Connected
NetSuite
Acme Corp — Production
Connected
3 integrations active · Last synced 6 min ago

Connects to Your Systems

Connect your expense platform, PoS, and ERP. The Data Agent pulls transactions from wherever your team spends, transforms the data against your rules, and pushes clean, coded entries into your accounting system. No spreadsheets. No manual reconciliation.

Your Transformation Rules, Applied Every Time

GL mapping, currency conversion, entity splitting, field formatting, and transaction categorisation. Captured clearly and applied consistently to every job. No more spreadsheet manipulation. No more copy-paste errors.

Data Agent/Rules/Expense Coding
1
Match each Revolut cardholder to a NetSuite vendor contact by name or email. If no exact match is found, flag the transaction for manual review.
2
Map expense categories to GL codes using the account mapping table. If a tag or note references a project code, override the default GL code with the relevant project cost account.
3
For software and digital subscriptions from EU suppliers, always apply reverse charge VAT at 23% regardless of what the invoice shows.
4
If the supplier is in the UK or outside the EU, code tax as zero-rated. Do not attempt to reclaim VAT shown on non-EU invoices.
5
For Irish suppliers, use the VAT rate shown on the receipt. If the expense category is travel, entertainment, or staff welfare, code tax as non-recoverable.
6
Use the absolute value of the Revolut amount. Format the reference as {merchant} {short_id} for traceability in NetSuite.
6 of 14 expense coding rules shown · View all →
Source Data
25/02 UBER EATS DUBLIN -€84.30
24/02 ADOBE CREATIVE CLOUD -€59.99
24/02 RYANAIR DUB-LHR -€187.40
23/02 AMAZON EU SARL -€214.88
23/02 COSTA COFFEE BAGGOT ST -€12.60
Revolut · 48 expenses
Transformed Output
Date Description GL Code Amount
25 Feb Uber Eats 6200 €84.30
24 Feb Adobe 6400 €59.99
24 Feb Ryanair 6100 €187.40
23 Feb Amazon 6300 €214.88
NetSuite CSV · Ready for import

Review Before Anything Exports

Every transformation job produces an output your team can review before it goes anywhere. Source data on the left. Clean output on the right.

Check the mapping. Verify the totals. Spot anything unusual. When it looks right, approve and export. Nothing leaves the system without meeting the controls you defined.

Built-in Validation, Every Job

Every transformation job runs a set of validation checks before the output is finalised. Totals balanced. No duplicates. All records mapped. Date ranges verified.

If anything fails, such as an unmapped transaction, balance mismatch, or missing field, it is flagged immediately with a clear explanation. Your team resolves exceptions before the data goes anywhere.

Job Summary — GL Journal Upload
Records Processed
48
Auto-mapped
98.2%
Flagged
6
Validation Checks
Debit/credit totals balanced ✓ Pass
All expenses mapped to GL codes ⚠ 6 flagged
No duplicate transactions ✓ Pass
Date range matches statement period ✓ Pass
Revolut · Feb 2026 expenses

Full Visibility. Full Audit Trail.

Every transformation explained. Every action logged.

Agent Reasoning — Expense Coding Explain
GL Code: 6200 — Meals & Entertainment
Revolut category is 'Food & Drink'. Merchant is Uber Eats. No project code found in tags. Rule: Map expense category to default GL code from the account mapping table. 'Food & Drink' maps to 6200 — Meals & Entertainment. Applied: 6200.
Tax Code: VAT on Purchases 23%
Supplier address on receipt is Dublin, Ireland. Category is Meals & Entertainment — not in the non-recoverable list (travel, entertainment, staff welfare). Rule: Irish supplier, standard category — use VAT rate from receipt. Receipt shows 23% VAT. Applied: VAT on Purchases 23%.
Contact: Emma Byrne
Revolut cardholder email is e.byrne@acmecorp.com. Matched to NetSuite contact 'Emma Byrne' by email address. Confidence: exact match.

See Why Every Transformation Was Made

For every field the agent maps, converts, or categorises, you can open the reasoning panel and see why. Which rule was applied, what source data was used, and how the output was derived. If something looks off, the full trail is there.

Full Audit Trail, Every Job

Every stage a transformation job passes through is logged. When it was triggered, what rules were applied, who reviewed the output, and when it was exported. Searchable, filterable, exportable.

Integrations
Rules
Jobs
Outputs
Logs
Timestamp Source Job Stage Actioned By
25 Feb 09:00 Revolut Expense Coding Triggered System
25 Feb 09:04 Revolut Expense Coding Transformed System
25 Feb 09:06 Revolut Expense Coding Reviewed sarah@acmecorp.com
25 Feb 09:07 Revolut Expense Coding Exported System
24 Feb 09:00 Revolut Expense Coding Triggered System
24 Feb 09:03 Revolut Expense Coding Transformed System
18 log entries today · Export →

See How Numra Can Automate Your Data Transformations.

Book a call and we'll show you the agent transforming real financial data.

Book a Demo →