Governance precedes deployment.
No model is connected to a workflow before its use case is registered, its data classified, and its oversight rules written.
Governance
The controls system in front of every model.
Veetso is operated by VEETSO LIMITED, a UK company that builds an AI-enabled financial platform. Governance is the part of the platform that decides what the technology is allowed to do. It arrives before the technology does.
Principles
Three statements that govern the rest.
Everything in the governance system follows from these. They are the rules a reviewer can hold against any proposed AI use case and decide quickly.
Intelligence supports people.
AI accelerates the work of staff, compliance officers, and operators. It does not own decisions that humans are accountable for.
Sources are attributed.
Internal answers carry source links. External drafting is marked. Nothing is presented as fact without a chain back to it.
The committee
Authority that is written down.
The AI Steering Committee is the body that decides whether an AI use case is allowed to proceed. Its terms of reference, membership, and meeting cadence are documented. Decisions are recorded against the use-case register.
Approves every use case
No model is connected to a workflow before the AI Steering Committee has reviewed and approved its registry entry, its data classification, and its human-oversight rule.
Owns the register
Maintains the use-case register as the single source of truth for what AI is doing inside the platform, who is accountable, and what stops it.
Sets acceptable-use rules
Publishes and updates the rules that distinguish allowed work from forbidden work, including external drafting, customer-facing surfaces, and any source of training data.
Reviews vendors
Approves every AI vendor before procurement, sets the no-training contractual basis, and re-reviews them on a documented cadence.
Answers to the board
Reports up to the directors of VEETSO LIMITED, the legally accountable officers of the company on the public UK Companies House filing.
Five pillars
The shape of the governance system.
Each pillar carries a set of documented controls. Together they cover the surface area a regulator would expect to see from a financial platform running AI workflows.
AI governance
Use-case register, acceptable-use rules, and a steering committee with documented authority.
Data protection
Classification, residency, encryption, retention, and access controls applied per use case.
Access control
Identity-bound permissions, source eligibility checks, and segregated environments.
Compliance support
Tooling for KYC, transaction monitoring, alert triage, policy drafting, and reporting.
Vendor oversight
Due diligence, security review, and ongoing monitoring of every AI technology partner.
Design controls
What the controls actually look like inside the product.
The same controls govern every workflow on the platform. Source attribution, access scoping, classification, oversight, and use-case separation are applied uniformly, not on a case-by-case basis.
Access control
Structured permissions, role-based access, and approved source controls govern every interaction.
Data classification
Information is classified by sensitivity, purpose, ownership, and AI eligibility.
Source awareness
Internal answers include source attribution. Nothing is presented without a chain back to it.
Human oversight
AI supports staff. It does not replace responsible human decision-making.
Use case separation
Internal knowledge answers and broader AI drafting are presented as distinct modes.
Further reading
The reasoning behind every gate.
Each gate has a short essay behind it, written by the team that built it. The collection is the responsible-AI series.