1. Connect
Connect the client’s accounting software and the other systems that support their financial operations, including AP, AR, and document sources.
Fabriq is the intelligence layer that works alongside your team—bringing together information from the ledger, AP, AR, email, documents, and spreadsheets.
Fabriq’s agents handle repetitive preparation: extracting data, matching transactions, drafting reconciliations, and preparing reports. Your people handle judgment, exceptions, and final approval.
The AI prepares the work. Your team stays in control.
The information needed to complete accounting work is often spread across the client’s accounting software, AP and AR tools, email, documents, and spreadsheets. Fabriq brings that information into one workbench so the firm can complete and review the work without repeatedly moving data between systems.
Connect the client’s accounting software and the other systems that support their financial operations, including AP, AR, and document sources.
Capture invoices and statements, match transactions, reconcile accounts, and prepare reports in one workbench.
Your team checks the proposed work. Only approved changes are written back to the client’s books.
Every client already has a general ledger — QuickBooks, Zoho Books, Sage. It is good at what it was built for: one company’s invoices, bills, and period close. That system of record should stay where it is.
A practice’s month is the work around that ledger. Staff pull invoices and bank statements out of PDFs, match each line to an account and a source document, then rebuild group reports the source GL cannot produce across entities and currencies. AP and AR often live in another system entirely. None of that is a “books” problem. It is a throughput problem — repeated across a book of clients on different software.
Replacing the client’s ERP is not realistic. A tool that only automates one slice (collections, spend, a single close checklist) still leaves extraction, reconciliation, and consolidation on the desk.
Fabriq connects to the client’s books and to their AP/AR tools, does the work the GL leaves behind, and writes approved changes back. The client keeps their software. The firm gets one workbench, with a person on every posting.
Select a system to see how Fabriq connects.
Then a human review inbox — nothing posts without sign-off.
A book of small clients on mixed QuickBooks and Zoho. The grind is capture, categorize, and rec — not a new chart of accounts.
Group clients, more than one entity, FX, a pack the source GL cannot produce cleanly. Reporting is the gap.
Bank and card volume that still gets matched line by line. Seniors review; juniors should not be hunting in two systems.
QuickBooks, Zoho Books, and Xero are live. Sage, Bill.com, and Salesforce follow the same bidirectional pattern and are on the roadmap.
Forward sync of invoices, payments, contacts, and ledger. Write-back on the reverse-sync pattern after review.
Same normalized model as QuickBooks. One workbench when the book of clients is mixed.
Direct sync into the same Fabriq ledger. Useful when a practice already spans QB and Xero.
Same bidirectional pattern as the live books connectors — not a second product.
AP documents and payments into the layer, so rec and reporting are not split from the bill workflow.
AR context (customers, invoices, collections state) next to the books — not a CRM replacement.
Multi-entity, foreign exchange, consolidation, and intercompany. Subsidiary books can stay in the client’s software; Fabriq is the HQ view, with drill-through to the posting.
Extraction maps into the client’s chart of accounts. A person reviews before anything is booked. This is the intern-in-Excel step, done in the same workbench as rec.
The agent matches bank lines to a counter-account and a source document, then queues proposals. A reviewer approves or rejects. Every decision is on the trail.
Templates today: book transactions, reconcile, categorise expenses, analyse spending, or a custom goal. The agent plans and proposes. Writes wait on approval unless the firm has chosen otherwise.
One workbench — three operational views, from staff through partner.
Start a Workbench session, drop statements, work the obvious matches. Ambiguous items stay in the inbox. The client’s QuickBooks or Zoho is not the place they spend the hour.
Approve, reject, or send back. The trail shows who proposed the match, what it was matched to, and who signed. Month-end review is a queue, not a reconstruction.
Standardize capture → rec → report without forcing a GL migration. Sell custom agents or a connector as a scoped engagement when a client does not fit the template.
Start with a 45-minute session on one mixed-software client. If it holds, we run a 30-day pilot on two or three clients: connect the books, import one statement cycle, run reconciliation and reporting in the workbench, and measure staff hours and unmatched volume against the month before.
A working session on your client stack.
What you can use today, and what follows on the same architecture.
| Capability | Status | Notes |
|---|---|---|
| QuickBooks, Zoho Books, Xero sync | Available | Bidirectional with the client’s books |
| PDF / image extraction | Available | Invoices, statements, receipts |
| Categorization and reconciliation with review inbox | Available | Human sign-off before posting |
| Workbench agent templates | Available | Book, reconcile, categorise, spending, custom |
| Multi-entity, FX, HQ consolidation view | Available | Group reporting across entities |
| Write-back to the source GL | Available | Approved changes sync back to the client’s books |
| Sage, Bill.com, Salesforce | Roadmap | Same bidirectional pattern as the live connectors |
| Full intercompany elimination close pack | Roadmap | Built on the existing multi-entity and FX model |