Approach
The AGORA platform
The tool AGORA uses to organise governance work during an engagement: which AI systems exist, how much risk they carry, which controls are met and with what evidence.
A tool of the engagement, not a separate product
Judgement comes from the AGORA team. The platform keeps what is decided, what is assessed and the evidence behind it, so the work does not depend on loose spreadsheets or on one person's memory.
It is used within an engagement, when the scope includes it. It is not sold separately or offered with open sign-up.
- It does not replace professional work or issue automatic conclusions.
- It does not certify or claim compliance: it organises the evidence for whoever must review it.
- It is not offered as a self-service product with public sign-up.
- Its use in an engagement is defined by scope, not by default.
What it does
Ten modules, in the order they are used
Each module supports one part of the journey. The stage shows where it is first used; it then stays in use through monitoring.
Which AI exists and how risky it is
The starting point: you cannot govern what you cannot see.
AI system inventory
Each system with its owner, its vendor and its lifecycle, from idea to retirement.
Stage 02 AssessRisk classification
A guided four-step questionnaire based on the EU AI Act —prohibited practices, Annex III, biometrics and transparency— and the equivalent reading under each framework that applies to the organisation.
Stage 03 Prioritise
How decisions are made and documented
Decisions are written down, with who approved them and when.
AIA, DPIA and FRIA impact assessments
Structured sections, a review and approval flow, expiry dates and PDF export.
Stage 04 GovernPolicies
Mandatory sections, published versions and a record of who read and signed each one.
Stage 04 GovernData protection
The sub-processor register, data subject requests with their deadline and retention of the audit log.
Stage 04 Govern
What is met and with what evidence
Governance is shown with evidence, not statements.
Controls and evidence
The status of each framework control, the evidence behind it —with a SHA-256 fingerprint to check it has not changed— and open findings.
Stage 05 OperationaliseModel cards
Performance and fairness of each model, versioned, with the difference between versions.
Stage 05 OperationaliseIncidents
A board by status, with deadlines by severity and each jurisdiction's regulator notification clock, with its template.
Stage 06 Monitor and scale
How leadership sees it
A view that makes sense without opening each system.
Leadership dashboard
Systems by risk level, control progress and upcoming deadlines. The G0–G3 maturity reading is confirmed in the review with AGORA, not automatically.
Stage 06 Monitor and scaleAudit log
Every change goes into a log that only accepts new entries: who did what and when.
Stage 06 Monitor and scale
What the dashboard looks like
A sample view, built with fictitious data to show what information it brings together.
- AI systems
- 12 in the inventory
- High risk
- 3 with an impact assessment
- Controls with evidence
- 58% of the agreed scope
- Deadlines
- 4 in the next 30 days
| System | Risk | Lifecycle | Owner |
|---|---|---|---|
| Inspection prioritisation | High | Pilot | Operations |
| Application assessment | High | Development | Risk |
| Service assistant | Limited | Production | Customer service |
| Document classifier | Minimal | Production | Records office |
Who uses it
Each person sees what concerns them. Access is granted by role and reviewed with the organisation.
- Data and AI leadership
- AI ethics
- Data protection
- Information security
- Model owners
- AI architecture and operations
- Legal and compliance
- Internal audit
- External audit, read-only and time-limited
Sensitive roles sign in with a second authentication factor.
Reference frameworks
The same frameworks as the AGORA framework. Risk classification is read under each framework that applies to the organisation.
International
- EU AI Act
- ISO/IEC 42001
- NIST AI RMF
- OECD AI Principles
- GDPR
National
- Chile · Law 21.719
- Peru · Law 31814
- Brazil · LGPD
- Mexico · LFPDPPP
Deadlines by jurisdiction are an input to legal review, not legal advice.
See it working
In a 30-minute conversation we show the platform with sample data and see whether it makes sense to include it in the scope.