Question 1

What kinds of documents does the AI read?

Accepted Answer

Receipts, invoices, and monthly statements — bank statements, credit-card statements, merchant statements, and Stripe/Shopify/Square payouts — plus the PDFs a client emails or texts at the last minute. Phone photos of crumpled receipts and forwarded email attachments all work; the extractor normalizes the layout and lifts the merchant, amount, date, and category into the transaction record.

Question 2

How does the anomaly flagging actually work?

Accepted Answer

Every posted transaction is run against a small ruleset tuned to solo-bookkeeper workflows: a recurring subscription that didn't bill this month, a category that doesn't match the merchant, an out-of-period posting from a statement that closed last cycle, and a handful of common data-entry tells (transposed digits, doubled decimals, a $0 tip that should have been $20). Anomalies are surfaced as a checklist on the client record so you can decide to chase the missing one or note it as a recurring variance.

Question 3

Will my clients see anything, or is this just for me?

Accepted Answer

Plumbmark is a back-office tool for one solo bookkeeper, not a client-facing portal. Your clients never log in; you collect their documents however you already do (email, text, a shared drive) and load them into the intake pipeline. That keeps the relationship on your side of the counter and avoids the awkward "did you get my email" text thread every month.

Question 4

How fast does a new client go from "I just signed the engagement" to "first reconciled month done"?

Accepted Answer

Once a client signs your engagement and uploads the opening balances along with one or two months of statements, most intakes finish in under ten minutes. The receipts and statements land as you receive them, the AI pulls the numbers, and the first month-end close walks out of the pipeline already reconciled. You're not waiting on the client to find last quarter's statements before you can bill.

Question 5

What happens to the data and the AI extraction — is my clients' information safe?

Accepted Answer

Documents are stored encrypted at rest and only your operator account can read them. The LLM extraction runs against a sealed extract-and-verify flow — the model sees the document for parsing only, never to train on, and the structured numbers are what get saved. No third-party retraining, no data sale, and the complete row is yours to export, archive, or delete at the end of the engagement.

Question 6

Is this a replacement for QuickBooks or my ledger of choice?

Accepted Answer

No. Plumbmark is the front-of-house intake and prep layer; it hands you a clean, categorized, anomaly-flagged file at the start of each month. You still post into the ledger you already know — the journal entries, the class tracking, the reconciliation tie-outs — Plumbmark's job is to make that hand-off fast and complete, not to replace the line-by-line work that defines your bookkeeper's value.

AI document collection for solo bookkeepers

Fifteen clients. Three deadlines. A receipt inbox that resets every Monday.

Three things that move the needle on a solo bookkeeper's month.

Solo bookkeepers ask, before they sign up.

Start the next month-end already on top of the receipts.