Skip to content

KaosKal/reclaimed-workforce

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

21 Commits
 
 
 
 
 
 
 
 

Repository files navigation

The Reclaimed Workforce: Digital Worker Reference Architecture

A governance framework for deploying AI Digital Workers in regulated financial services environments. Built for organisations operating under FCA, DORA, EU AI Act, and SMCR obligations.


What this repository contains

This repository is the technical companion to The Reclaimed Workforce four-part series published on LinkedIn, with the Champion's Guide publishing on Medium.

It provides the reference architecture, governance templates, and regulatory mapping for deploying Digital Workers with cryptographic identity, policy enforcement, and a tamper-evident audit trail.

The series addresses a straightforward problem. Skilled people in financial services spend the majority of their working week on high-volume, process-intensive work that consumes their expertise without rewarding it. Digital Workers can handle that work. But an ungoverned AI agent operating inside a regulated firm is not a productivity tool. This framework makes the difference.

The Reclaimed Workforce — LinkedIn series:

Champion's Guide — publishing on Medium. Link to follow.


The core problem this solves

An AI agent without governance is a liability, not an asset.

No identity. No policy. No audit trail. No accountability.

This framework gives every Digital Worker:

  • A unique, time-limited cryptographic identity (SPIFFE/SVID)
  • A Policy Decision Point that enforces its Digital Job Description before any action executes
  • An append-only Attribution Ledger with cryptographic chain of custody
  • A tested Kill Switch with a maximum 15-minute revocation window
  • A named human accountability owner on every record

Repository structure

reclaimed-workforce/
├── README.md                  # This file
├── LICENSE                    # CC BY 4.0
├── diagrams/                  # Architecture diagrams (SVG)
│   ├── vocabulary-bridge.svg
│   └── from-wish-to-mandate.svg
├── governance/                # Governance templates
│   ├── digital-job-description-template.md
│   └── five-gates-checklist.md
└── regulatory-mapping/        # Regulatory alignment tables
    └── eu-ai-act-mapping.md

The Five Gates

No Digital Worker moves from Shadow Mode to Full Agency until all five gates are cleared. Each gate has a named owner and a defined evidence requirement.

# Gate Owner Evidence required
1 Identity Gate Business Owner SPIRE registration confirmed. SVID issued and validated.
2 Policy Gate Risk & Compliance Signed DJD approved. OPA bundle loaded and verified.
3 Shadow Gate Technical Lead Shadow Mode log reviewed. Kill Switch tested.
4 Audit Gate GRC Analyst Ledger records present. Chain integrity verified.
5 Conformity Gate CISO / DPO DPIA completed. Conformity assessment documented.

A worker that clears four gates is not almost ready. It is not ready.


Key components

Digital Passport (SPIFFE/SVID)
A time-limited cryptographic identity for each worker. Not a shared service account. Standard TTL: 15 minutes. High-assurance: 5 minutes.

Policy Bus / Policy Decision Point (OPA)
Every action checked against the approved Digital Job Description before execution. Deny-by-default posture.

Attribution Ledger
HMAC-chained append-only audit log. Cryptographic chain of custody. Satisfies EU AI Act Article 12 logging requirements.

Digital Job Description
The governance document that defines what the worker may do, expressed as a signed OPA policy bundle. Satisfies EU AI Act Article 14 human oversight requirements.

Kill Switch
SPIRE registration entry deletion. Revokes the SVID within the TTL window. Must be tested before go-live. A Kill Switch that has never been tested is not a control.


Regulatory alignment

This architecture is designed for compliance with:

  • EU AI Act — Articles 12 (logging), 14 (human oversight), 17 (quality management)
  • DORA — ICT risk management and incident response
  • FCA Consumer Duty — Audit trail and accountability requirements
  • SMCR — Named individual accountability for every worker
  • NIST AI 600-1 — Shadow Mode satisfies pre-deployment safety evaluation requirements for high-risk agentic systems

Full regulatory mapping table in /regulatory-mapping/.


Status

Component Status
Reference architecture diagrams Available
Digital Job Description template Available
Five Gates checklist Available
EU AI Act mapping table Available
Implementation examples Planned

Author

Kal Perwaz — AI Security Architect and GRC Specialist
30 years in infrastructure and governance. Financial services.
LinkedIn · GitHub


Licence

This architecture and documentation is published under the Creative Commons Attribution 4.0 International (CC BY 4.0).
You are free to use, adapt, and build on it. Attribution required.

About

AI governance framework for financial services — The Reclaimed Workforce series by Kal Perwaz

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors