See what actually exists
Read-only inventory of sources, volumes, formats, risks, duplicates, encryption and context.
A new 2dA product family for digital-born data
Digital data to archival standards and knowledge for every next flow
BronVast brings existing digital data to agreed archival standards for integrity, formats, metadata, context, rights and sustainable storage. The knowledge created in that process can become rules, evaluation sets and task-specific models that also support other and future data flows.
2dA BronVast
Digital data to archival standards
Fixity, format validation, metadata, rights, preservation events and controlled information packages make the data sustainably manageable.
Processing knowledge becomes reusable
Taxonomies, examples, evaluation sets, rules and local models can also support new data flows.
The route from source to lasting organisational capability
Know the sources
media, fileshares, mailboxes and exports
Control the data
integrity, format, privacy and relations
Build context
metadata, provenance and responsible roles
Capture knowledge
rules, examples, embeddings and task models
Support new intake
recognise, describe, route and execute under control
Practice view 01
A shared drive containing fifteen years of project data
BronVast registers the source, checks integrity and formats, reconstructs structure and context, and delivers files with metadata and processing evidence through the agreed archival route.
Archival first, then reusable elsewhere
The first result is digital data demonstrably brought to agreed archival standards. The second is the controlled knowledge layer created during that processing.
- Scan & Intake for a reliable baseline
- Formats & Preservation for sustainable technical control
- Metadata & Context for findability and explanation
- Local Intelligence for reusable rules, evaluation sets and models
- BronVast Flow to apply that knowledge in other data flows
What remains besides the collection
The engagement delivers more than files. It creates a controlled knowledge layer the organisation can continue to use.
- source registers and checksum manifests
- format policies, decision rules and exception logs
- metadata profiles, vocabularies and taxonomies
- approved training and evaluation examples
- model cards, confidence thresholds and feedback history
What the organisation retains
The backlog becomes governable
Volumes, risks, cost drivers and routes are visible before major migration begins.
Knowledge becomes transferable capability
Decision rules, metadata, examples, evaluation sets and models remain documented rather than becoming one-off project knowledge.
Future flows do not restart at zero
A reusable model can recognise document types, propose metadata, run checks and guide new intake to the right route.
Practice view 02
The same knowledge then supports new intake
Document profiles, approved examples and exceptions from the project can support a local model. On a new upload it can recognise type, extract fields, propose metadata and recommend the right process route.
Local by design
The entire chain can run locally
Source files, metadata, embeddings, vector store, model inference and suitable fine-tuning can remain on customer infrastructure or in a segregated 2dA environment. Hardware, licences, access, logging and retention are defined per route.
01Source zone02Processing zone03Metadata and vector layer04Local models05Chat and workflows
Technical reference points
We align the route with recognised standards and specifications where they fit the source, destination and customer responsibility. This is never presented as automatic certification.
PRONOM and DROIDFormat identification using signatures and PUIDs, not extensions alone.PREMISObjects, events, agents and rights as a preservation metadata foundation.E-ARK and OAISWhere appropriate, align with SIP, AIP and DIP concepts and interoperable packages.NIST AI RMFGovernance, evaluation, documentation and monitoring for local AI.
Questions about this route
Must we use all BronVast modules at once?
No. An engagement can start with a read-only scan or one concrete source flow. Routes work separately and can be connected later.
Can BronVast operate fully locally?
Yes, when infrastructure and selected models support it. Data, metadata, embeddings, vector store and inference can remain local. The architecture is agreed per engagement.
What are reusable models?
Task-specific models or configurations built from controlled examples, rules and corrections created during processing. They can later classify document types, extract fields, propose metadata, test quality or route information in other data flows. Versions, tests and human corrections remain documented.
Can such a model later operate inside an existing data flow?
Yes. After evaluation, a model can run locally beside a mailbox, upload service, DMS, case system or data export. It can make proposals and perform permitted routine tasks as data arrives. Destruction, appraisal, rights and other high-risk decisions remain with an authorised user.
Next step
Choose the right starting point
Start with one source, one process or a read-only assessment. We will define what should remain local, what evidence is required and which route creates value first.