The problem they faced
A growing accounting firm serving 80+ small business clients was stuck in what one team member called "Stone Age" workflows. Despite having bank feeds configured in their cloud accounting platform, staff still manually entered invoice lines, cash deposits, and credit card expenses — employees submitted paper cover sheets with GL codes that were then typed in by hand. Month-end consolidation across multi-entity clients required exporting trial balances, rearranging columns in spreadsheets, running VLOOKUPs on account codes, and hunting for differences when figures didn't balance. One senior accountant spent 8 hours per month just on consolidation for a single client group. The firm was turning away new clients because they literally didn't have the manpower — a common problem among accountants who are too busy typing to take on new clients.
What Phoenix AI built
Phoenix AI automated the firm's core data entry and consolidation workflows. Bank transactions are now processed with AI-assisted categorisation: the system suggests codes based on the firm's own historical coding patterns, and a staff member reviews and approves each suggestion before it's applied. Unusual or low-confidence transactions are flagged for manual coding, keeping the firm in full control of what gets posted. Receipt images submitted by clients are processed through OCR, extracting dates, amounts, vendors, and tax — then matched to bank transactions automatically. Month-end consolidation runs as a scheduled workflow: trial balances are pulled from each client's system, account codes are mapped and normalised, and a consolidated workbook is generated with variance highlighting. The firm also implemented automated client follow-up for missing documentation, so staff no longer spend time chasing receipts and bank statements.
Outcomes delivered
The firm saved an estimated 20 hours per week across the team on manual data entry alone. Month-end consolidation time was cut in half — from a multi-day process to a few hours of review. The firm accepted 12 new clients in the quarter following implementation, revenue they would have had to turn away. The team finally stopped complaining about data entry — accountants could focus on advisory work and client strategy instead of "typing numbers from one screen to another."