Framework
The framework provides five core components: the financial semantic foundation, the semantic runtime, governed business workflows, open interfaces, and AI-Control. Each is introduced in the sections below.
Click the diagram to enlarge
AI-Control operates across the other components, providing governance, policy, permissions, validation, monitoring, compliance alignment and runtime assurance. Explore Finsight AI-Control →
Financial Semantic Foundation
Give financial data connected, machine-readable meaning. Traditional schemas describe how data is stored—they rarely explain what it means, how financial concepts relate or which rules govern its interpretation. The Financial Semantic Foundation captures this context in a structured form that people, applications and AI can consistently understand and apply.
Industry-Aligned Foundation
Starts from established financial concepts, modelling patterns and industry standards—the ISDA Common Domain Model, the Financial Industry Business Ontology, regulatory data models and common product and lifecycle concepts. Standards provide the common vocabulary; they do not replace your terminology or operating model.
Finsight Semantic Repository
Reusable semantic structures and implementation assets: semantic metamodels, financial ontologies, canonical concepts, relationship definitions, mapping structures, rules and control patterns, evidence relationships and reference artefacts—so you do not design every structure from the beginning.
Institution Semantic Repository
Each institution extends the common foundation with its own business terminology, product and process models, systems, schemas, confirmed mappings, transformation rules, lineage, ownership, policies and operational knowledge. Not a generic ontology—an institution-specific foundation built from reusable components.
AI-Foundry Semantic Runtime
Make the semantic foundation operational. Rather than treating an ontology as static documentation, the runtime compiles, validates, queries and applies semantic knowledge during execution.
Repository Loader & Compiler
Loads and validates semantic artefacts, resolves references and relationships, and produces a consistent, fingerprinted repository snapshot.
Graph Runtime
Represents concepts, data assets, mappings, rules, controls and evidence as a queryable semantic graph—traversal, discovery and impact analysis.
Semantic & Hybrid Retrieval
Connects user language to relevant context—combining semantic search, graph traversal and structured filters, not just similar text.
Evidence & Provenance
Captures the evidence behind every result—sources, derivations, temporal validity, confidence—so outputs can be explained and traced.
Lineage & Temporal Context
Connects technical data movement with business meaning: where data originated, what transformed it, which version was effective and what is affected.
Capability Kernel
Reusable, governed functions that workflows, applications and AI agents invoke—separating foundation capabilities from workflow-specific logic.
Workflow Orchestration
Coordinates retrieval, data access, evidence assembly, AI capabilities and validation—with execution policies, failure handling and auditability.
Governed AI Integration
AI models are accessed through structured, controlled runtime interfaces—bounded context, validated outputs, provider abstraction and telemetry.
The runtime provides the reusable technical foundation. Institutions focus on their critical business data, logic and controls.
Plug-in Business Workflows
Apply the semantic foundation to real business outcomes. Plug-in Business Workflows combine deterministic execution with bounded AI, evidence, validation and human oversight to deliver governed, traceable and explainable results.
Deterministic workflow logic
controls the process
Semantic capabilities
provide business and data context
AI
supports selected reasoning tasks
Evidence and controls
govern the result
One reusable foundation. Multiple controlled, business-specific workflows.
Interfaces and Open Integration
Make semantic and governed capabilities available across the institution. AI-Foundry operates as an open capability layer rather than a closed user application.
Who can use it
- Business applications
- Internal AI assistants
- Enterprise AI agents
- Data and analytics tools
- Orchestration platforms and workflow engines
- Developer applications
How it is accessed
- REST APIs
- Model Context Protocol (MCP) tools
- Command-line interfaces
- Python interfaces
- Web applications
- Embedded application components
Other AI agents can use AI-Foundry to resolve financial concepts, discover institutional data, retrieve graph-grounded context, trace lineage, obtain approved evidence and invoke governed workflows—reusing one financial semantic foundation across many AI applications instead of embedding meaning into every assistant.
Open to your applications and agents. Controlled by defined contracts and policies.
AI-Control
Governance and runtime assurance across the architecture. AI-Control is loosely coupled and complements the institution's existing governance, risk, security and compliance frameworks.
Safety and governance are first-class architectural capabilities—not controls added after implementation.