
Finance Automation Strategy for an Independent School
- Client
- Independent school and nursery · London · around 200 children
- Industry
- Education
- Department
- Finance & Accounting
- Region
- United Kingdom · UK & Ireland
- Duration
- 3 weeks
A London school and nursery ran its entire finance function through one bursar, and the bursar was leaving. Every invoice, direct debit, bank line and supplier bill passed through one person's hands and one person's memory. We mapped every process, designed the automation that connects their systems, and gave the directors a plan for month-end to survive the resignation letter.
A working model of the approach on sample data. It plays on its own and pauses when scrolled away; nothing here touches a live system.
The results
- 20 h → 1 h
- weekly finance work Projected
- Projected: review and exceptions only.
- 780 h
- returned a year Projected
- Projected, payback in about a year.
- 200
- children billed and reconciled every term
- Across two systems with no usable API.
The challenge
Our client is an independent pre-prep school and nursery in London with around 200 children and 45 staff. Modern tools were already in place, but isolated: Connect Childcare for sessions and parent billing, Pegasus for accounting, GoCardless for direct debits, Barclays for banking, spreadsheets for everything in between. The bursar was the manual integration layer between all of them.
What it was costing them
| Area | How it worked |
|---|---|
| Parent billing | About 180 invoices generated, checked and sent one by one each term; every session change recalculated and resent |
| Collections | Direct debit plans loaded by hand, the bank checked monthly, payments matched manually, non-payers chased individually |
| Supplier bills | 50+ invoices a month collected from email, keyed in, reconciled and paid manually |
| Statutory calendar | HMRC, Teachers' Pension, quarterly VAT and year-end certificates on a manual calendar |
- A single point of failure. If the bursar was ill, on holiday or leaving, finance stopped. And the bursar was leaving.
- A growth ceiling. Fifty more families meant more bursar hours, not a configuration change.
- Connect Childcare → The bursar
- Pegasus → The bursar
- GoCardless → The bursar
- Barclays → The bursar
- Spreadsheets → The bursar
The approach
We delivered a paid digital transformation assessment, an automation strategy and a full solution design, with a costed build plan the directors could act on.
Assessment: what could be automated, and how much
Current-state income and expenditure were mapped step by step, with as-is, as-is by cadence and to-be process maps. Eight opportunities were scored for potential, effort and value:
| Opportunity | Automatable |
|---|---|
| Migration to Xero (foundation) | Enabler |
| Accounts payable | 80% |
| Parent invoicing | 85% |
| Payment monitoring and collections | 90% |
| Parent self-service portal | 75% of queries |
| Teachers' Pension and payroll | 65% |
| Grant funding optimisation | 60% |
| Supplier contract and renewal management | 70% |
- Term billing & grants
- Direct debit cycle
- Bank reconciliation
- Supplier invoices
- Human review queue
Strategy: a five-phase, 12-month roadmap
- Foundation. Pegasus to Xero, with bank feed, vendor history import and native integrations, recommended alongside the school's own advisors.
- Expenditure. Accounts payable automation.
- Income. Billing and reconciliation.
- Self-service. A parent portal.
- Intelligence. Forecasting, governance reporting and late-payment prediction.
The core build we designed
- Term billing and grants. Child records and funding allocations updated, grant confirmations loaded, eligible children selected for the funding period, exceptions flagged for a person rather than assumed.
- Direct debit cycle. Billing generated and exported, extra items added, the collection file uploaded and validated before anything is charged.
- Reconciliation engine. Bank lines matched per child, returned direct debits handled, unmatched items routed to review. The accounting package's API does not expose reconciliation and its bank rules do not scale to 200 families, so matching is done programmatically.
- Supplier invoices. Bills read by document AI and mapped to account, cost centre and VAT treatment from the school's 70-vendor matrix, then entered for review.
- Systems with no integration surface, driven anyway. Connect Childcare through our Python and Robot Framework desktop engine; Xero through our Playwright-based nodes in n8n; n8n orchestrating the whole cycle.
- Controls. A human review queue, run reports and alerts, and one supervised parallel month alongside the bursar before anything runs on its own.
- Connect Childcare → n8n and review queue
- Xero → n8n and review queue
- GoCardless → n8n and review queue
- Bank → n8n and review queue
Technology
| Layer | What we designed with |
|---|---|
| Orchestration | n8n |
| Web automation | Our Playwright-based n8n nodes (Xero) |
| Desktop automation | Python and Robot Framework (Connect Childcare) |
| Documents | Document AI and OCR for supplier bills |
| Finance stack | Xero, GoCardless, open banking feed, DocuSign, Pegasus to Xero migration |
The outcome
All figures below are projected from the assessment and the school's own numbers.
| Today | With the automation (projected) | |
|---|---|---|
| Finance work a week | ~20 hours | ~1 hour of validation |
| Finance work a month | ~70 hours | ~5 hours, review and exceptions |
| Hours returned a year | ~780 | |
| Who month-end depends on | One bursar | A documented, repeatable process |
| 50 more families | More bursar hours | A configuration change |
- About 20 hours a week of finance work reduced to 1 hour of validation, roughly 780 hours a year returned.
- Payback on the build in about a year.
- Month-end no longer depends on one person: documented, repeatable and auditable.
- A full-time bursar role becomes a part-time coordinator handling exceptions, approvals and planning.
- Growth becomes a configuration change, not a hiring decision.
The risk most teams miss
Key-person risk hides behind a process that works. It only shows the week that person leaves, and by then there is nobody left to document it.
Built with
The tools this ran on, by the layer each one carried.
- Apps & integration
- AI
- Automation
Next step
Start the same conversation
Tell us the process that keeps breaking. We will come back with what it would take to fix it, what it would cost, and what we would measure.
Or write to us directly: •••••@•••••••.•••
- A reply from a partner within one business day
- Fixed scope and fixed fee, agreed before we start
- The maps and the backlog are yours either way





