The Universal_FSCompliance_MCP Project's mission is to enable Financial Institutions to make it easier to bring the right financial products to consumers safely and cost-effectively.
Compliance with Standards is an unavoidable and necessary part of being a Financial Institution. These Standards have grown to be compendious. This is a natural habitat for AI, in which LLMs take seconds to ingest half-million-word documents.
Standards are words-based, and therefore native to LLMs (whose middle initial is "Language"). AI-native Standards Compliance Tools thus have an inherent advantage.
With only pre-AI tools, or where AI is added later rather than being core to the architecture, compliance operations can be cumbersome and expensive, leading to higher costs and a drag on innovation, leading ultimately to poorer outcomes for consumers.
Our AI-agentic Tools are AI-native. Accessing them from their AI agents will bring about a radical reduction in friction and cost for Financial Institutions.
The Tools are not intended to be autonomous, and human judgement will still be required. But the Tools will do much of the heavy lifting. Users, typically compliance officers and other professionals working in Financial Institutions, will be able to focus on senior executive tasks.
The Tools will be held in a suite of specialist MCP servers. (MCP, or Model Context Protocol, is an open framework published by Anthropic, and is widely used in the construction of AI Agents). Each MCP server will cover an Implemented Standard. The ambition is universal in the sense of being capable of being applied to any codified Standard. Our first Implemented Standard is the FCA Handbook.
The MCP architecture is important. MCP servers can easily be built into AI agents, without lengthy integration work. AI agents are set to become increasingly prevalent in process management in Financial Institutions, so MCP architecture will ease implementation and allow easy changes to processes. This is a significant advantage over custom APIs.
The MCP servers are available in an open-source repository on GitHub and a permissive MIT license. They are designed for enterprise self-hosting.
Accuracy and reliability have been engineered into Tool specification and operation. Financial Institutions' security and risk management requirements are embedded at the core. Scaling and performance have been designed in.
This repo contains overarching governance and management documents. Separate repos are maintained for USP Servers for each of the Implemented Standards.
-
Visualisation: Interface mockup showing the Universal FSCompliance MCP's broad capabilities.
-
FAQ.md: User questions, platform capabilities, implementation timelines, and enterprise adoption requirements
-
SpecialFactors.md: AI Agents, Tools, MCP Architecture, Standards, Ground Truth.
-
ComplianceTools.md: MCP tool specifications, implementation roadmap, and Ground Truth validation system
-
LLMChoice.md: AI model selection, multi-provider architecture rationale, and enterprise flexibility strategy
-
Differentiators.md: Competitive advantages, market positioning, and investment thesis for funding evaluation
-
OurStory.md: Founder narrative, development journey, and strategic insights.
-
Touchstones.md: List of factors that bind the documentation together.
-
LegalLiability.md: Legal framework, liability limitations, user responsibility.
-
Charging.md: Enterprise-grade monetization framework with configuration-driven billing architecture and transparent value-based pricing, initially set at zero.
-
DatabaseArchitecture.md: Database architecture for holding Standards and MI in a way suited to AI Tool use.
-
TechnologyStack.md: System design, infrastructure requirements, security specifications, and deployment architecture.
The Universal_FSCompliance_MCP Project is opensource, licensed under the MIT License. The MIT License is a permissive license that allows free use, copy, modification, and distribution of software without restriction.
About This Document
Author: Blake Dempster, Founder, CEO, Principal Architect
Last Updated: 16 August 2025