Procupy Lens reads the exports from your accounting system, runs about 35 standard checks and writes a plain-language memo on what's wrong, what changed and where the money is. Every figure in it traces back to your data.
$1.9M on 493 invoices, 84% of everything you're owed.
2
Travel costs as much as payroll
High
$377K against $374K of payroll, 20% of revenue.
3
Approved expense claims never paid
Medium
239 claims, about $90K, waiting a median of 500 days.
Each figure is checked against the computed facts before the memo is shown.
Lens on a public sample ledger pack, Jan 2023 to Jun 2025
What you upload
Start from the exports you already have
Lens works from files, so there is nothing to connect. Upload Excel (.xlsx, .xlsm, .xls) or CSV files of up to 50 MB each, one at a time or the whole month-end folder.
Exports are rarely tidy. Lens finds the header row on its own, works out which kind of file each one is and maps the columns. When it isn't sure what a column means, it asks you instead of guessing.
File
What Lens looks at
General ledger
Revenue, cost of sales and operating costs by period, margins, unusual postings
Vendor bills (payables)
Open and overdue bills, late payments, reliance on a few suppliers
Customer invoices (receivables)
Overdue invoices, late payers, reliance on a few customers
Budget vs actual
Overspend by department and period
Expense claims
Approved claims never paid, rejection rates, large claims that got through
How it works
From a folder of exports to a memo in one sitting
01
Upload
Drop in a ledger, bills, invoices, a budget and expense claims. Any mix works.
02
Confirm the columns
Lens maps every column and only asks about the ones it can't place with confidence.
03
Lens runs the checks
Plain calculations on your data, the same way every time. No AI model decides a number.
04
Read the memo, then ask
Findings come in money order. Ask follow-up questions in plain words and see where each answer came from.
What it checks
About 35 checks a finance director would run
Every finding has a code, a severity and the numbers behind it, so the same problem reads the same way every month.
Data quality
A ledger that doesn't balance, currencies mixed without exchange rates, placeholder descriptions, posting patterns that don't look like normal bookkeeping.
Profit and loss
Margins shrinking year on year, revenue falling while costs rise, strong seasonality, cost lines out of line with revenue.
Payables and receivables
The share of bills or invoices still open, how many were paid late and by how many days, reliance on one supplier or customer.
Budgets
Which department overspent and by how much, and the largest single-period swings.
Reconciliation
Invoices, bills, budget actuals or expense claims that don't agree with the ledger.
Anomalies
Monthly spikes, outliers within an account, rare combinations of account, department and counterparty.
Opportunities
The biggest costs you can actually control, and spend spread across many vendors that is worth putting out to competitive bidding.
The checks are plain calculations, not guesses by an AI model. Lens then writes a short memo in plain language from those computed facts and checks every figure in it against the evidence. Numbers it can't trace are rejected.
The memo is written from the computed facts, not from your raw ledger rows. If the model is switched off, you still get a rules-based memo, the findings and every table.
Ask
Ask your books a question
Ask in plain words. Lens answers by querying the data you uploaded, shows which query each number came from and marks any figure it couldn't verify.
Which vendors did we pay late last quarter?
How did travel spend move against budget?
Which customers owe us the most past 90 days?
Month after month
Several companies, one view
Keep separate books for each entity and compare them side by side in a group view. A new analysis can carry forward the files from the last one, skip duplicate rows and show what changed since the previous run, so a monthly review becomes an upload and a read.
Who it's for
Built for people who read the books for a living
Finance controllers and CFOs doing a monthly or quarterly review without a BI team.
Accountants, advisers and auditors who receive a folder of exports from a client and need a first read quickly.
Procurement heads who want a clear picture of spend before a sourcing round.
Groups with several entities that want one view across them.
Part of Procupy
Lens uses the same account as Procupy, the procurement platform. When Lens shows a category spread across many vendors, you can put it out to an RFQ or a reverse auction from the same workspace.
Procupy Lens turns general ledger, payables, receivables, budget and expense-claim exports into a list of findings, a plain-language memo and a chat you can ask questions. It is part of Procupy and uses the same account.
Which files does Lens accept?+
Excel (.xlsx, .xlsm, .xls) and CSV exports of up to 50 MB each. Lens finds the header row and maps the columns on its own, and asks you when a column is ambiguous.
Do I need to connect my accounting software?+
No. Lens works from exported files, so it works with any accounting system that can export to Excel or CSV.
Can the AI make up numbers?+
The checks are calculated directly from your data, not by an AI model. The model only writes the memo and answers questions, and Lens checks the figures it uses against your data and rejects or flags anything it can't trace.
Is Lens free?+
Yes. Each workspace gets 20 analyses and 200 Ask questions a month at no cost. Workspaces on a paid Procupy plan have no Lens limits.
Is it the same account as Procupy?+
Yes. Sign in with your Procupy account, or create one here. The same login works on procupy.com, and being signed in on one site signs you in on the other.
See what your books say
Create a free workspace, upload this month's exports and read the memo.