Trusted AI Digital Repository

Build AI on knowledge you control.

Bring together your collections, metadata, rights, and institutional knowledge in a governed environment so you can create AI-powered experiences without giving up control of your data, context, or policies.

Illuminated manuscript leaf from a royal psalter
Illuminated manuscript leaf from a royal psalter
Bowl decorated with a central rosette
Bowl decorated with a central rosette
Hand-drawn map from a sketchbook by Daniel Huntington
Hand-drawn map from a sketchbook by Daniel Huntington

Research

Education

Public access

Visitor experiences

Illuminated manuscript leaf from a royal psalter
Bowl decorated with a central rosette
Photograph of adults and children on a village street in the Auvergne
Hand-drawn map from a sketchbook by Daniel Huntington
Letter to Eugรจne Maus decorated with two painted apples

Research

Education

Public access

Visitor experiences

Images: The Metropolitan Museum of Art, Public Domain โ€” Psalter leaf ยท Bowl ยท Village street ยท Sketchbook map ยท Letter to Eugรจne Maus

Your AI should work from your knowledge

Your institution already has the knowledge AI needs.

Your collections, digital assets, archival material, metadata, rights information, provenance, and institutional expertise already form a body of trusted knowledge.

A Trusted AI Digital Repository connects those sources in a governed environment so you decide what AI can access, how it can use that information, and which experiences it can support.

Uncontrolled information

Draws from the open web

Unclear provenance

May include outdated/inaccurate information

No control over sources

Higher risk of errors/misattribution

Trusted institutional knowledge

Verified collections/records

Rich metadata/context

Rights and permissions applied

Governed access/use

Grounded in institutional expertise

A Terentia foundational framework

The Five Layers of Trust

Trust is not a single feature. It has to be designed across the knowledge environment.

01

01

Trusted Assets

Trusted Assets

Authoritative files, versions, derivatives, preservation masters, access copies, and relationships.

Authoritative files, versions, derivatives, preservation masters, access copies, and relationships.

02

02

Trusted Metadata

Trusted Metadata

People, places, events, creators, provenance, themes, cultural meaning, and relationships.

People, places, events, creators, provenance, themes, cultural meaning, and relationships.

03

03

Trusted Rights

Trusted Rights

Access rules, licensing, restrictions, sensitivities, permissions, and reuse logic.

Access rules, licensing, restrictions, sensitivities, permissions, and reuse logic.

04

04

Trusted Workflows

Trusted Workflows

Human review, approvals, auditability, stewardship practice, and cross-team accountability.

Human review, approvals, auditability, stewardship practice, and cross-team accountability.

05

05

Trusted Intelligence

Trusted Intelligence

AI grounded in institutional knowledge rather than uncontrolled internet content.

AI grounded in institutional knowledge rather than uncontrolled internet content.

The framework stays consistent. The implementation is yours. Your systems, policies, workflows, collections, and AI priorities determine how the five layers come together in your environment.

How it comes together

Start with what you already have.

Start with what you already have.

A Trusted AI Digital Repository helps you connect the knowledge that already exists across your institution and make it usable in a controlled, governed way.

A Trusted AI Digital Repository helps you connect the knowledge that already exists across your institution and make it usable in a controlled, governed way.

1. Curate your trusted sources

Determine which collections, repositories, systems, metadata fields, and assets belong in the trusted knowledge environment.

2. Build AI contexts

Create purpose-specific contexts with the right approved data, permissions, instructions, and levels of human review.

3. Deliver many experiences

Support research, education, public discovery, visitor engagement, collections workflows, and internal knowledge from one governed foundation.

Platform integrations

Work with the environment you already have.

A Trusted AI Digital Repository does not have to begin with a rip-and-replace project. Connect existing collections systems, DAMS, archives, research databases, shared repositories, and other institutional sources through APIs and integrations. The goal is to create a governed knowledge layer around the environment you already operate.

One trusted foundation. Many possibilities.

What could you build on top of your trusted knowledge?

Handwritten letter from W. H. Fox Talbot to Antonio Bertoloni, ink on paper

Collections workflows

Use AI to support OCR, transcription, metadata enrichment, image and video indexing, and discovery while retaining human review.

Hand-drawn map from a sketchbook by Daniel Huntington

Research & discovery

Enable natural-language and semantic search across trusted collections so researchers can uncover records, relationships, and primary sources.

Floral design by Antoine Berjon, a study of flowers

Public access

Create more intuitive ways for audiences to explore collections while maintaining control over the content and context being surfaced.

Cactus plant in flower

Education

Build purpose-specific learning experiences grounded in approved institutional sources.

Bowl decorated with a central rosette

Visitor experiences

Create conversational, personalized, or interpretive experiences that remain connected to institutional context.

Diagram: collections systems, digital assets and shared drives connected through a Trusted AI Digital Repository so staff can find institutional knowledge in one place.

Internal knowledge

Help staff find and use institutional knowledge distributed across systems, departments, and repositories.

Your institution. Your implementation.

Build around your data, systems, and priorities.

Build around your data, systems, and priorities.

There is no single AI architecture that works for every institution. Your environment may include a DAMS, collections management system, archive, research database, shared drives, or legacy tools. The Five Layers of Trust provide the framework. How they are applied should reflect your institution.

There is no single AI architecture that works for every institution. Your environment may include a DAMS, collections management system, archive, research database, shared drives, or legacy tools. The Five Layers of Trust provide the framework. How they are applied should reflect your institution.

01 ยท Trusted Assets

What knowledge do we want AI to use?

02 ยท Trusted Metadata

Which sources do we trust?

03 ยท Trusted Rights

What should remain restricted?

04 ยท Trusted Workflows

Where should human review happen?

05 ยท Trusted Intelligence

Which experiences matter most to our audiences and staff?

A real-world example

See how another institution approached it.

The Theodore Roosevelt Presidential Library faced its own version of the challenge: historical material distributed across collections and institutions, with ambitious plans for research, discovery, and public engagement.

Terentia serves as the Trusted AI Digital Repository for TRPLโ€™s collections environment, providing a governed foundation from which downstream research and public experiences can draw.

Campfire, the Theodore Roosevelt Presidential Libraryโ€™s AI research assistant, answering a question with cited collection sources.
Campfire, the Theodore Roosevelt Presidential Libraryโ€™s AI research assistant, answering a question with cited collection sources.

Resources

Continue exploring trusted AI

Letโ€™s talk about your environment

Build AI on your terms.

You may already have the right collections, systems, and expertise. We can help you look at your current environment, identify the knowledge that matters, and explore what a Trusted AI Digital Repository could look like for your institution.

Your data.

Your environment.

Your rules.

Your AI.

FAQs

A governed institutional knowledge environment that connects collections, digital assets, metadata, rights, permissions, and contextual knowledge so they can safely support AI-powered experiences.

No. A DAMS or CMS may be an important source within the environment, but the Trusted AI Digital Repository is the broader governed knowledge layer that can connect multiple systems and repositories.

Not necessarily. The goal is to work with your existing environment where possible, connect trusted sources, and make approved information available to AI in a controlled way.

No. Different collections, records, fields, and assets can have different permissions. Different AI experiences can also receive access to different subsets of institutional knowledge.

Yes. A public discovery tool, internal assistant, collections workflow, and education experience can each have different approved datasets, rules, permissions, and purposes.

Start with the institutional challenge you want to solve. Then identify the trusted sources, governance requirements, workflows, and experiences needed to address it.

About the imagery on this page

Collection images shown on this page are used as illustrative examples of cultural heritage content and are sourced from public-domain or open-access collections. Their inclusion does not imply any commercial relationship, endorsement, sponsorship, or partnership between Terentia and the contributing institutions. The Theodore Roosevelt Presidential Library is featured separately as a Terentia customer and case study.