A global education and heritage organisation
Entity-linked genealogy records across census, parish and migration archives, with handwritten-text recognition
Published 24 July 2026
At a glance
| Industry | Education and heritage | genealogy and family-history programmes |
|---|---|---|
| Region | Global | delivery across 40+ centres in 50+ languages |
| Record types | 3 | census · parish registers · migration records |
| Core capabilities | 3 | genealogy data curation · structured family-history datasets · heritage archival digitisation |
| Hardest technical problem | Entity linkage | resolving one person across archives under variant spellings, scripts and decades |
| OCR scope | Handwritten-text recognition | materially harder than printed OCR — historical hands vary by scribe, era and region |
| Quality standard | 95%+ accuracy SLA | same standard and dual-layer review as every Lifewood programme |
| Status | Continuing partnership | complemented by the Malaysia Genealogy Event activation, May 2026 |
This is a Lifewood Data Technology case study in Genealogy · Heritage — an engagement delivered for a global education and heritage organisation across Global · Education. Entity-linked genealogy records across census, parish and migration archives, with handwritten-text recognition
Scanning the page is the easy half; resolving the identity across archives is the work
A global education and heritage organisation required entity-linked data across census, parish and migration records, with archival OCR including handwritten-text recognition. The difficulty is not capture. A census page, a parish register and a migration record may all refer to the same person under different spellings, in different scripts, recorded decades apart — and the value of the dataset is entirely in connecting them correctly. Scanning the pages is the easy half; resolving the identities across them is the work. Handwritten-text recognition compounds it. Historical hands vary by scribe, era and region, and the source documents are often damaged, faded, or bound in ways that resist flat scanning. This is a materially different problem from printed OCR, and it is the reason archival digitisation programmes stall at the image-capture stage rather than producing searchable records.
A long-running partnership producing entity-linked records, not page images
Lifewood maintains a long-running data partnership with the organisation, delivered across 40+ delivery centres and 50+ languages, covering genealogy data curation, structured family-history datasets and heritage archival digitisation. The deliverable is a structure rather than a scan: a person, their recorded events, and the documentary sources that evidence each one, joined so a family line can be traced across archives. 1. Archival capture including handwritten-text recognition. Applied across historical hands that vary by scribe, era and region, on documents that are frequently damaged or bound. 2. Entity resolution across record sets. Variant spellings, differing scripts and decades of separation are reconciled to a single identity, which is where the dataset's value actually sits. 3. Structured family-history dataset construction. Records are output as linked people, events and sources rather than as document images. 4. Dual-layer review under the standard SLA. Held to the same 95%+ accuracy standard as every other Lifewood programme, across the 40+ centre network. That structure is what makes an archive searchable and, increasingly, what makes it readable by machines rather than only by researchers.
A continuing partnership anchoring Lifewood's heritage and genealogy capability
The partnership continues across genealogy data curation, structured family-history datasets and heritage archival digitisation, at the standard 95%+ accuracy threshold. Entity-linkage quality is assessed on correct resolution of the same individual across separate record sets rather than on character-level OCR accuracy alone.
Related Lifewood services
Verified outcomes
| Metric | Value | Baseline | How measured |
|---|---|---|---|
| Accuracy standard | 95%+ | contractual SLA across all Lifewood programmes | Dual-layer human-in-the-loop review |
| Record types linked | 3 | census, parish and migration reconciled to one identity | Entity-linked output structure |
| Delivery footprint | 40+ centres | across 50+ languages | Lifewood centre network |
| Output form | Entity-linked records | rather than page images | Structured family-history datasets |
Method and verification. Figures on this page are Lifewood-reported. Accuracy is measured against a customer-approved gold set under dual-layer human-in-the-loop review at the standard 95%+ threshold. Entity-linkage quality is assessed on correct resolution of the same individual across separate record sets rather than on character-level OCR accuracy alone, because a perfectly transcribed page linked to the wrong person is a worse outcome than a partially transcribed page linked correctly. The partnering organisation is not named on this page under the confidentiality terms of the engagement.
Questions about this programme
On Lifewood genealogy and heritage programmes the hardest part is entity linkage — resolving one person across census, parish and migration records recorded decades apart.
Yes — archival OCR on this heritage digitisation partnership includes handwritten-text recognition, which is a materially different problem from printed OCR.
This Lifewood genealogy and heritage partnership covers genealogy data curation, structured family-history datasets and heritage archival digitisation, under the same 95%+ accuracy SLA as every other programme.
On Lifewood heritage data programmes the deliverable is entity-linked records rather than page images — a person, their recorded events, and the sources evidencing each one.
Run a similar program with Lifewood?
Tell us your scope and accuracy bar. We will scope a comparable engagement within one call.
Talk to our team