Every accounting firm knows this pattern: client documents arrive constantly — by email, by portal upload, sometimes still by scan — with no consistent naming, no consistent format, and no indication of which client folder they belong in until someone opens them. At busy periods, sorting and filing alone can eat hours a day that should be going toward actual client work.
We build automation that removes that sorting step entirely: documents are identified, classified, and filed into the correct client folder automatically as they arrive.
The problem in practice
A typical firm receives a mix of invoices, receipts, bank statements, tax correspondence, and identity documents from dozens of clients, often in a single shared inbox or upload portal. None of it arrives labelled. Someone has to open each file, work out what it is, work out which client it belongs to, and file it in the right place — a task that's repetitive, easy to get wrong under time pressure, and doesn't require anyone's actual expertise to do.
Misfiled or unfiled documents aren't just an inconvenience. They cause missed deadlines, duplicate requests to clients for paperwork that was already sent, and audit trail gaps when something needs to be located months later.
What we actually build
The automation we build for this follows a consistent pattern:
- Intake — a flow watches a shared inbox, portal, or SharePoint upload folder for new documents
- Classification — Azure Document Intelligence and Azure OpenAI read the document and identify what it is (invoice, receipt, bank statement, tax form, identity document) from its content, not just its filename
- Client matching — the document is matched to the correct client using details already in the document (name, ABN, account reference) against the firm's client records
- Filing — the document is renamed consistently and filed directly into that client's folder in SharePoint, in the right sub-category
- Exception handling — anything the system can't confidently classify or match is flagged for a person to review, rather than filed incorrectly or silently dropped
The result is a client folder structure that's always current, without anyone manually maintaining it.
Why this needs to be built properly, not bolted on
The hard part isn't reading a single clean PDF — it's handling the real mix firms actually receive: low-quality scans, photos of receipts taken on a phone, multiple documents merged into one PDF, or documents that reference more than one client entity. A workflow that only handles the clean cases ends up creating more manual cleanup than it saves. Getting the exception handling right — routing anything ambiguous to a person instead of guessing — is what makes this reliable enough to run unattended during a firm's busiest periods.
What this solves for an accounting firm specifically
- Time back during peak periods — tax time and BAS deadlines are when document volume is highest and admin capacity is lowest; automated filing removes exactly the bottleneck that hurts most
- A genuine audit trail — every document's origin and filing decision is logged, which matters for compliance reviews
- Faster turnaround for clients — work can start as soon as a document is filed and matched, not once someone gets around to sorting the inbox
- Fewer "can you resend that" requests — nothing gets lost in an unsorted inbox
If your practice is still sorting client documents by hand, a free automation assessment will show exactly what's automatable in your current intake process — book a time and we'll walk through it.