Data Catalogue

Connect enterprise data to AI without giving up control.

T-Flux Ultra gives customers granular control over enterprise data at every stage of the AI lifecycle.

Data can be version-controlled, governed, and managed from ingestion through to retrieval and model use.

Granular data governance

Keep every source, version, and retrieval path under enterprise control.

Organisations can determine which version of a document or data set is current and authorised for AI use; when a new version is introduced, approved, superseded, or retired; and who may access information by user, role, tenant, agent, or model.

Policies can govern classification, retention, indexing, embedding, retrieval, exposure to a model, human approval, and audit trails at each processing stage.

Data sources and types

Connect the information your organisation already uses.

Documents and enterprise files

PDF, DOCX, XLSX, and uploaded artefacts

Connect reports, policies, contracts, spreadsheets, manuals, procedures, research material, and other enterprise documents through a governed ingestion pipeline.

Enterprise repositories

SharePoint, DMS, Box, S3, and document repositories

Keep existing repositories as the authoritative source while T-Flux creates the indexes, metadata, and retrieval capability required for enterprise AI.

Email and knowledge sources

Email archives, knowledge bases, and wikis

Incorporate approved email repositories, internal knowledge bases, operating procedures, wikis, and institutional knowledge into a governed retrieval environment.

Databases and APIs

Structured data, operational systems, and API feeds

Connect current structured information from databases, applications, and API-enabled systems while retaining existing application and data boundaries.

Manual upload

Controlled ad hoc ingestion

Authorised users can add approved information for temporary use, testing, analysis, or inclusion in a governed knowledge corpus.

AI-ready knowledge

From enterprise information to governed AI context.

Connecting information is only the first step. T-Flux prepares enterprise information so it can be reliably retrieved and used by approved AI models.

01
Connect
Bring approved enterprise sources into the governed environment.
02
Parse and extract
Process formats and use OCR where needed to make content usable.
03
Classify and structure
Add source, ownership, classification, and context metadata.
04
Index
Create vector and keyword indexes for efficient discovery.
05
Retrieve
Provide the most relevant authorised information as contextual evidence.

This is the foundation of Retrieval-Augmented Generation. T-Flux can ground AI responses in the organisation's own authoritative information rather than relying solely on the information contained inside a large language model.

Data flow from enterprise sources through the T-Flux Data Plane and governance controls to the T-Flux Model Catalogue.

Data sovereignty

Your data remains your data.

Corporate data remains under corporate control. T-Flux does not require an organisation to relinquish ownership of its information in order to use AI.

Data can remain within on-premises infrastructure, private cloud, approved data centres, controlled enterprise repositories, and jurisdictional boundaries.

T-Flux operates as the orchestration and governance layer between enterprise data, applications, users, and AI models. The organisation retains control over what information can be accessed, who can access it, which models can use it, and for what purpose.

Governance before AI access

A permissions layer sits between enterprise information and the AI environment.

Access Control Lists

Synchronise existing repository permissions so users and AI workloads retrieve only information they are entitled to access.

Role and attribute-based access

Apply RBAC and ABAC policies by role, business unit, security level, application, or other customer-defined attributes.

Trust classification

Assign trust labels that determine how and where information can be used.

Jurisdiction controls

Tag information by geographic or regulatory jurisdiction to support sovereign and regulatory requirements.

Retention policies

Keep information subject to enterprise retention and lifecycle requirements.

Sensitive information controls

Identify, mask, or redact PII and protected information according to customer-defined policy.

Data-aware multi-model orchestration

Give every approved model the right context for the request.

The Model Catalogue provides access to multiple AI models within the same governed environment. The Data Catalogue and Data Plane provide the approved enterprise context those models are allowed to use.

Customer-defined policies can determine which model handles a request, which enterprise information it is permitted to see, which sources are relevant, which guardrails apply, and how the resulting response is validated and recorded.

Operational visibility

Monitor every connected source through the Data Plane.

The T-Flux Ultra Data Plane provides operational visibility over the complete enterprise AI information lifecycle, including connected sources, ingestion volumes, document processing, OCR and parsing success, indexing, embedding creation, vector database utilisation, retrieval performance, failures and retries, access-control enforcement, knowledge-source utilisation, and data lineage.

Explore the Data Plane