Some of what it costs you
never reaches your books.
LZR checks your bank, invoices, GST and payroll against each other and catches it before your return locks.
You find out what June looked like in August. By then the duplicate is a write-off and the 2B mismatch is a lapsed credit., the 2B mismatch is a lapsed credit, and both are corrections rather than decisions.
Days between period end and books you can act on.
Anomalies caught before close rather than after.
Three steps, and a person stands in the middle of the second one.
Bank statements, invoices, GST returns, TDS records and payroll arrive. Each entry gets a proposed category, a confidence figure, and the reasoning that produced it.
Approve, change, or flag. Nothing posts to the books on the agent's word alone, and the reviewer can disagree with it.
Every entry carries the person, the rule, or the agent that set it, and the time it happened. The trail is readable later without reconstruction.
A flag states what was observed, names the rule that raised it, and offers exclusive outcomes.
Dismissing a flag is a recorded decision and asks for a reason. It is not a close button. Violet appears here and nowhere else in the product: it means a person needs to look at this. Red is kept for the system failing, never for your money.
- REASON
- A fixed reason code, not free text.
- CLAIM
- One or two lines stating what was observed, never what to do about it.
- EVIDENCE
- The two records side by side (invoice against bank line, books against GSTR-2B) and the fields they differ on.
- SOURCE
- Which rule, agent, or person raised it, and when. A flag with no attributable source is not shown.
- RESOLVE
- Two or three mutually exclusive outcomes, one of which is acceptance.
A flag is not a tidiness note. It is two documents disagreeing about money that already moved or money that never arrived.
A ₹40,000 purchase sits in your books. Your supplier never filed it, so it isn't in your GSTR-2B. That's ₹7,200 of input credit you cannot claim.
Nothing in your own records looks wrong. The gap exists only between two documents and it repeats every month.
First payment ever to this account. ₹85,000, and the creditors ledger has no prior history with this vendor. It gets a second look before it clears, not after.
Changed bank details on a familiar invoice is how most of this money leaves. The stakes are theft, not tidiness.
The same vendor invoice is booked twice in creditors and paid twice from the bank, in the same week. Flagged before the second one clears, not found three months later.
Recoverable at the bank this week. A credit note you chase for a quarter after that.
An invoice crosses the e-invoice limit, and your filed GSTR-1 has no e-invoice against it. That's a compliance gap, not a line item.
Missed, it becomes a notice. Nobody runs a query to find this one and it looks correct until someone asks.
This vendor's bill is usually ₹12,000 a month. This month it's ₹38,000, no explanation. Flagged, not paid on autopilot.
A question asked before the payment goes out, while the vendor still has to answer it.
Form 26AS shows ₹1,20,000 of TDS credited against you. Your debtors ledger accounts for ₹96,000 of it. The ₹24,000 flagged before the return is filed.
Either revenue you never invoiced or credit you're about to forfeit. No reconciliation run surfaces this on its own.
A flag is a proposal, not an action. Nothing is paid, held, or filed because LZR said so — your accountant approves it, changes it, or rejects it, and the ledger records who decided. The same six shapes run across every other pair: GSTR-1 against 3B, salary slips and PF against the bank, Form 16A against the TDS you have already claimed.
The limits are the product. Four of them are worth stating before anything else.
It does not file anything and it does not sign anything.
Filing and sign-off stay where they already are, with the firm and the person qualified to do them. LZR prepares and evidences the review that comes before that.
It does not post an entry nobody accepted.
A proposal at 99% confidence and a proposal at 40% are both proposals. Neither reaches the books until a reviewer accepts it.
It does not overwrite a correction.
When a reviewer changes a category, the agent does not change it back on the next run. The correction becomes the rule for that payee.
It does not ask you to trust a number you cannot open.
Every proposal shows what it matched on. Every flag names the rule that raised it. A conclusion with no visible reasoning behind it is not shown at all.
This is the first question every firm asks and the page cannot answer it yet. The confirmed list of supported ledgers goes here before launch. A firm reading a vague answer to this question will assume the answer is no.
Four things that change for you.
Speed to close
Days between the end of a period and books you can make a decision on.
Visibility
Categorised, current numbers instead of a three-month-old summary, without emailing your accountant to ask where things stand.
Money caught before it settles
The six shapes above (read across bank statements, invoices, GSTR-1, 2B and 3B, Form 16A and 26AS, payroll, creditors and debtors) surface while the period is open and the money is still recoverable.
No added headcount
The review happens inside your accountant's practice. You do not hire an internal finance person to get any of the above.
Send this page to whoever keeps your books.
LZR reaches a business through its accounting practice. If yours already uses it, ask to be shown the ledger view for your own last period. If they don't, the button below writes the message for you.
Neither button creates an account. There is no self-serve signup, because there is no way for a business to buy LZR on its own.