The Challenge
Volunteer-driven intake. Regulator-grade reporting. Both, every day.
Donations arrived in every imaginable format — paper forms scanned at regional offices, online submissions, bank transfers, gift-aid declarations from estates. Each had to be classified, validated against gift-aid rules, reconciled against the donor record, and surfaced in a finance system within audit windows.
The volunteer network that kept it running was being asked to absorb more administrative burden every quarter, and a single misclassification could surface in a Charity Commission inspection two years later.
The Approach
Take volunteers out of the audit trail — not out of the mission.
Three phases, each scoped around the donor experience and the regulator's expectations.
Document intake
OCR and NLP models trained on five years of historical donation forms across formats. Confidence-scored extraction with human review only for ambiguous fields, not every record.
Compliance encoding
HMRC gift-aid rules, safeguarding flags, and donor-consent policies encoded as a versioned rule engine. Auditors receive a complete chain from policy to classification on demand.
Donor unification
Three legacy fundraising platforms unified behind a single donor record without forced migration. Volunteers keep their tools; leadership gets one number.
Continuous improvement
Quarterly model retraining anchored in volunteer feedback and regulator queries. Qubiz remains the embedded engineering partner for the platform.
What We Built
Document intelligence pipeline
An OCR and NLP pipeline that ingests donation forms, gift-aid declarations, and bank statements in a dozen formats — extracting structured fields and reconciling them against the donor record without human routing.
Regulatory rule engine
A versioned rule engine encoding HMRC gift-aid eligibility, charity-commission reporting requirements, and internal safeguarding flags. Every classification is auditable and every change traceable to the policy that drove it.
Donor system unification
Connected three legacy fundraising platforms and a finance ERP into a single donor view — without forcing migration. Volunteers and finance teams keep the tools they trust while leadership gets one source of truth.
Volunteer-first rollout
Trained over 200 regional volunteers on the new intake flow before launch. Adoption was measured in queue depth, error rates, and volunteer feedback — not user counts.
The Results
A platform that scales with the mission, not against it.
- 3.2M donations processed annually through the unified intake pipeline.
- 94% reduction in volunteer hours spent on classification and data entry.
- 0 material findings in Charity Commission audits since go-live.
- 200+ regional offices migrated onto the platform with no service disruption.
In Their Words
“Qubiz built something that respects how our volunteers actually work. The technology disappeared into the background, and our regional teams could go back to focusing on the people we're here to serve.”
Margaret Whitfield — Director of Digital, Healthcare Netherlands
Frequently asked questions
01How does the intake pipeline handle so many formats?
OCR and NLP models trained on five years of historical donation forms extract structured fields with confidence scoring. Human review is reserved only for ambiguous fields, not every record.
02How is regulatory compliance guaranteed?
HMRC gift-aid rules, safeguarding flags and donor-consent policies are encoded as a versioned rule engine, so every classification is auditable and traceable to the policy that drove it.
03Did volunteers have to change the tools they use?
No. Three legacy fundraising platforms were unified behind a single donor record without forced migration, so volunteers keep their tools while leadership gets one source of truth.
04What impact did it have?
The platform processes 3.2M donations a year, cut volunteer classification and data-entry hours by 94%, and has had zero material findings in Charity Commission audits since go-live.
