Ask three archivists what “retroconversion” means and you will likely get three confident, correct answers, and one nod towards a filing cabinet nobody has opened in years. The term is old, precise, and quietly essential. It is also routinely mixed up with something else entirely.
What does retroconversion mean?
Retroconversion — short for retrospective conversion — is the process of turning existing catalogue records into structured digital records. Not the documents. The catalogue.
If your finding aid lives on index cards, in a typed list from 1987, or in a handwritten accession register, retroconversion is the work of getting that descriptive information into a searchable digital system, properly structured, with consistent fields.
The term has been standard usage in library and archive science for decades, well before “digital transformation” entered anyone’s vocabulary. Librarians used it through the 1980s and 1990s to describe converting card catalogues into early automated systems (MARC records, mostly). Archivists inherited the term for the same problem applied to finding aids: paper description of collections, converted into digital description of collections.
Retroconversion is not digitisation
This is the distinction that trips people up, so it is worth stating plainly: digitisation is about the source material. Retroconversion is about the catalogue.
Digitisation means scanning or photographing the documents, photographs, maps or objects in your collection so a digital surrogate exists. We have covered that process in detail elsewhere, including why digitisation on its own is not the finish line if the resulting images sit undescribed.
Retroconversion means converting the description — the catalogue record that says what the item is, where it sits in the collection, and how it relates to everything around it — from paper into digital form.
Crucially, these two things can exist entirely independently of each other. A collection can be fully catalogued on cards with no image ever scanned. A collection can be extensively digitised with no structured catalogue behind the images at all. Retroconversion addresses the first situation specifically: a finding aid that is accurate, often excellent, and completely invisible outside the room it is kept in.
Who actually needs this
Three situations come up again and again.
Special collections and community archives with no archival management system. A local history society, a cathedral archive, a company heritage collection — often built up by dedicated volunteers or a single archivist over many years, with description recorded faithfully on cards or in ring binders, and no database anywhere in sight.
County record offices with card indexes that predate any database. Many local authority archives have decades of accession-level or item-level description on cards, created long before their current catalogue system existed, and never brought into it.
Institutions inheriting an unstructured paper accession register. A new archivist arrives, or a collection transfers between organisations, and the only record of what was acquired, when, and from whom is a paper log with no consistent structure to convert cleanly.
In every case, the description already exists. The problem is not a knowledge gap. It is a format gap.
Why it matters for discoverability
A paper finding aid, however thorough, cannot be searched by a researcher in another city, indexed by a search engine, or cited as evidence in a funding bid. It cannot be queried, cross-referenced, or exported to a national aggregator such as Archives Hub or Discovery.
The description on the card might be excellent. It does not matter. If it only exists on the card, the collection is functionally invisible to anyone who does not walk into the search room and ask.
This is the archival version of “crap in, crap out” — except the input is not crap at all. The problem is that good description trapped in an inaccessible format produces the same outcome as no description: a collection nobody can find, funders cannot see evidence of, and researchers assume does not exist. Retroconversion is what closes that gap.
What a retroconversion project actually involves
The shape of the work is fairly consistent, whatever the starting format.
Assess the existing finding aid. What structure does it already have, even informally? Cards are often ordered by accession number or by series; a typed list may already group items by fonds or collection. This existing logic is worth preserving, not discarding.
Decide a mapping to a standard. Most retroconversion projects map the existing description onto ISAD(G)-aligned fields — fonds, series, file, item — so the resulting catalogue is structured consistently and exports cleanly to other systems later.
Key or extract the description. Historically this meant manual rekeying, card by card. It can now be accelerated with AI-assisted transcription and extraction that reads the source cards or lists and drafts structured records for review.
Structure into a hierarchy. Individual records get organised into the fonds/series/file/item structure agreed at the mapping stage, so relationships between items are preserved rather than lost as a flat list.
Quality review. Every converted record needs checking against the original by someone who understands the collection, catching mis-transcriptions and inconsistencies before anything is treated as final.
Export. Once reviewed, records move into whatever system the institution will use going forward, in a standard export format.
How Archivers.ai helps
Archivers.ai’s managed retroconversion support is built around exactly this workflow. It combines AI-assisted transcription and extraction from your existing cards, lists or registers with structuring into an ISAD(G)-aligned fonds/series/file/item hierarchy, so the result is a properly organised catalogue rather than a pile of loose records.
Every converted record carries a draft or review status, with evidence cited on every AI-generated field, per-word OCR confidence on transcribed text, and full traceability back to the source card or list it came from. Nothing reaches a public portal until a person has checked it — retroconversion accelerates the mechanical work, it does not remove the professional judgement that decides whether a record is right.
If your collection has no digital catalogue system at all, Archivers.ai’s managed retroconversion can take you from paper straight into a structured, exportable catalogue, with onboarding and field-mapping support built in. If you already have a system and are moving away from it rather than starting from paper, that is a different project — see our guide to migrating from CALM instead.