We built a fictitious £30m charity as a comprehensive test case — 185 staff, 37 cost centres, 3 fund types, restricted and designated funds — and ran a full automation assessment against realistic finance data to prove what's possible.
The full test case data is available on request — GL transaction listing, chart of accounts, cost centres, payroll reports, and month-end timetable. We're happy to walk you through the working.
VGS is a fictitious £30m UK charity created as a realistic test environment. It mirrors the complexity of a real mid-to-large charity: 185 staff across 7 directorates, 37 cost centres, restricted and designated funds, an endowment, and a finance team of 8 FTE.
| Annual income | £30m |
| Staff | 185 |
| Finance team | 8 FTE |
| Cost centres | 37 across 7 directorates |
| Fund types | Unrestricted, Restricted, Designated, Endowment |
| GL account codes | 93 |
| Year-end | 31 March |
| Finance system | Sage / SunSystems |
| Payroll | Outsourced to bureau |
Before automation, VGS operates a 10 working day close at Level 1 automation maturity — documented processes, the ERP used as a data store, and reconciliations in Excel. This is consistent with APQC benchmarks for a charity of this size.
| Phase | Days | Focus | Hours |
|---|---|---|---|
| Pre-close | WD−5 to WD0 | Cut-off reminders, recurring templates, suspense clearance | 14.5 |
| WD1 | Day 1 | Bank feeds, purchase/sales ledger close, 21 credit card statements | 10.25 |
| WD2 | Day 2 | Payroll journal (185 staff, 37 CCs, 3 funds), accruals, depreciation, deferred income | 13.0 |
| WD3 | Day 3 | 4 bank accounts, debtors, creditors, payroll control, donation platforms | 9.5 |
| WD4 | Day 4 | VAT, fixed assets, prepayments, accruals, 8 restricted funds, 4 designated funds | 9.0 |
| WD5 | Day 5 | Analytical review of 93 GL codes, correction journals, VAT return | 9.0 |
| WD6 | Day 6 | FC sign-off, system lock, data extraction | 4.25 |
| WD7 | Day 7 | SOFA by fund, balance sheet, variance tables, 7 directorate reports | 11.5 |
| WD8 | Day 8 | Executive summary, variance commentary, 7 budget holder notes | 11.0 |
| WD9 | Day 9 | FD review, incorporate comments, quality sign-off | 6.25 |
| WD10 | Day 10 | Distribute to SLT, trustees, budget holders; close retrospective | 3.5 |
| Total monthly close effort | 101.75 hrs | ||
With AI Finance Office automation deployed, the timetable compresses to 4–5 working days. Total effort drops from 102 hours to 29.5 hours per month — a 71% reduction. The remaining hours are predominantly review, judgement, and approval tasks that require human expertise.
| Phase | Before | After | Reduction | What changes |
|---|---|---|---|---|
| Pre-close | 14.5 | 6.0 | 59% | Recurring templates auto-updated, suspense auto-cleared |
| WD1 | 10.25 | 4.0 | 61% | Bank feeds auto-imported, credit cards auto-coded |
| WD2 | 13.0 | 3.5 | 73% | Payroll auto-split, accruals auto-calculated, depreciation auto-posted |
| WD3 | 9.5 | 1.5 | 84% | Bank rec 90%+ auto-matched, sub-ledger recs automated |
| WD4 | 9.0 | 1.5 | 83% | Balance sheet recs auto-prepared, fund balances auto-checked |
| WD5 | 9.0 | 2.0 | 78% | Trial balance auto-reviewed, corrections auto-suggested |
| WD6 | 4.25 | 1.0 | 76% | Data extraction automated; period lock remains manual |
| WD7 | 11.5 | 2.0 | 83% | Management accounts pack auto-generated |
| WD8 | 11.0 | 3.0 | 73% | Variance commentary auto-drafted; human review only |
| WD9 | 6.25 | 3.0 | 52% | FD review unchanged — this is a judgement task |
| WD10 | 3.5 | 2.0 | 43% | Distribution automated; retrospective remains human |
| Total | 101.75 | 29.5 | 71% |
The automated close freed the equivalent of 3 months of full-time capacity annually. That freed capacity was redirected from data processing to the work that actually matters: better analytics for budget holders, timely narrative for the executive, and board-quality reporting delivered days earlier than before.
The finance team moved from chasing numbers to explaining them.
Individual processing time reductions across 18 month-end automations.
| Automation | Current (hrs/mth) | Automated | Saving |
|---|---|---|---|
| Bank Reconciliation | 2 | 12 min | 90% |
| Month-End Accruals | 4 | 8 min | 97% |
| Correction Journals | 3 | 15 min | 92% |
| Prepayment Journal | 2 | 10 min | 92% |
| Deferred Income Journal | 3 | 15 min | 92% |
| Fixed Assets & Depreciation | 1.5 | 5 min | 94% |
| Payroll Journal & Variance | 1.5 | 6 min | 93% |
| Income Reconciliation | 4 | 20 min | 92% |
| Budget Consolidation | 4 | 10 min | 96% |
| Management Accounts | 4 | 30 min | 88% |
| Cash Flow Forecasting | 2 | 15 min | 88% |
| Balance Sheet Forecasting | 2 | 15 min | 88% |
| Account Reconciliation | 2 | 12 min | 90% |
| Journal Consolidation | 1 | 5 min | 92% |
| Donor Reporting | 3 | 20 min | 89% |
| Journal Entry Analysis | 1 | 10 min | 83% |
| Credit Card Expenses | 1.5 | 8 min | 91% |
| VAT Return | 2 | 15 min | 88% |
| Total | 44.5 hrs | 3 hrs 46 min | 91.5% |
90% of finance teams are at Level 0–1. Moving to Level 3 is achievable with AI Finance Office support.
| Level | Name | Close Cycle | VGS Position |
|---|---|---|---|
| 0 | Manual | WD15–20+ | Below current state |
| 1 | Standardised | WD10–15 | CURRENT — documented processes, ERP as data store, Excel reconciliations |
| 2 | Partially Automated | WD7–10 | After Phase 1 — bank feeds, auto-reversing accruals, templated journals |
| 3 | Substantially Automated | WD4–7 | TARGET — AI reconciliations (>90% match), predictive accruals, auto-generated reports |
| 4 | Intelligent / Autonomous | WD2–4 | Stretch target — continuous close, anomaly detection, predictive models |
The time savings are significant — but the real transformation is in what the finance team does with the freed capacity:
Better analytics for budget holders. Instead of receiving numbers a week late, budget holders get variance reports with narrative commentary within days of month-end. They can act on the information while it's still relevant.
Timely reporting to the executive. The CEO and SLT receive management accounts with insight, not just data. The FD's time shifts from producing the pack to interpreting it and providing strategic advice.
Board-quality reporting, delivered earlier. Trustees receive a complete finance paper — SOFA, balance sheet, cash flow forecast, reserves analysis, KPI dashboard, and written commentary — with time to read it before the meeting.
A finance team that's upskilled and engaged. The team learns to work with AI as a professional tool. That's a career skill that makes them more valuable, more engaged, and more likely to stay. Organisations that invest in their finance team's development retain their best people.
Audit readiness as a by-product. Every automated journal has a full audit trail. Every reconciliation is documented. The year-end audit becomes a review of well-organised records rather than a scramble to reconstruct working papers.
The results in this case study reflect a complete transformation — every automation deployed, every process redesigned, full governance documentation, and comprehensive team training. That's our Enterprise package. But you don't have to start there.
| Package | What you'd get from this assessment | Monthly saving | Price |
|---|---|---|---|
| Rapid Assessment | The assessment itself — full process audit, automation opportunity register, and prioritised roadmap. This is where every engagement starts. | — | £1,500 |
| Foundation | 2–3 core automations built and tested — typically bank reconciliation, accruals, and correction journals. Team training, workflow documentation, and 30-day support. | ~15 hrs | £6,500 |
| Transformation | Full month-end journals, reconciliations, and reporting — 5–8 automations with process redesign, AI governance documentation, comprehensive training, and 60-day support. | ~32 hrs | £12,500 |
| Enterprise | Everything shown in this case study — the full transformation across all automations, multi-department rollout, custom development, complete AI governance and committee reporting, change management, and 90-day support with quarterly reviews. | ~40+ hrs | Bespoke |
The Rapid Assessment fee is credited in full when you proceed to Foundation or above. All automations run on Claude, Anthropic's AI platform — a Claude subscription is required, and we'll advise on the right plan for your team during the engagement.
Send us your GL transaction export and your month-end close timetable. We'll show you exactly what's possible for your organisation.