FairMind
Request access
Menu

0items approved by you24,000 checks in the run · 0 changed without you

Agentic AI governance

Governance that runs itself.
Decisions that stay human.

FairMind runs governance as agent workflows: collect, check, suggest, act, verify, file. Every workflow stops at a gate, and a person decides. Right now, that person is you.

Request access How a run works

What you just did

Every tile is a check in a run. It stayed grey until you looked at it.

Deny by default. Nothing is recorded or changed until an approver decides, and every decision is kept with who made it and why.

Behind you

Your decisions become the evidence trail.

A hash-chained trace of inputs, model calls, tool calls, and approvals, mapped to the requirements it satisfies. Auditors read the trail; they don't take our word for it.

51Frameworks in the registry, EU AI Act to OWASP Agentic
90Governance workflows agents can run
109Evaluations mapped to requirement IDs
21Packs across Govern, Evaluate, Secure, Comply
0External changes made without a human approval

How a run works

The agent stops at you. Nothing moves until you decide.

One real workflow from the catalogue, the deployment release gate, taken apart on a bench. Every station reads its step definition. Press one to look inside. The line backs up at the gate until a person decides.

Example runDeployment release gate and staged rolloutrun · deployment_release_gate · A2 Propose

At the gate · AI system owner + Risk

The tour never presses this. A person does.

0 filed by you · ledger head none yet

  1. Start the run Starts when the pipeline requests approval.
  2. Collect prerequisite status Reads status and dates of required workflows for this release.
  3. Check prerequisites Blocks if any prerequisite is missing, stale or failed; checks rollout plan and rollback path.
  4. Suggest decision and cohorts Proposes go / no-go and staged rollout plan.
  5. Dry-run the change Simulates the proposed change, computes blast radius, confirms rollback path and that credentials are scoped to this change only.
  6. Ask the AI system owner + Risk Sends AI system owner + Risk a decision card for Deployment release gate and staged rollout: what would change, why, the risk if approved, the agent's confidence band and links to the evidence.
  7. Release approval token Posts the signed approval to the CI/CD gate.
  8. Verify the change Re-reads target state and confirms each approved change applied and nothing else changed; on mismatch rolls back and escalates.
  9. File evidence Files prerequisite checklist, decision, token, rollout plan.

Autonomy

You set how far the agent goes. Per workflow.

Internal audit stays at Assist. Inventory upkeep can run at Act. External systems are never written without a person.

A0ManualThe agent does nothing unless asked for a draft.
A1AssistCollects and checks. Humans write every conclusion.
A2ProposeProposes; nothing is recorded until a human approves. The default.
A3Act on internal recordsFiles records inside FairMind; humans review by sampling. Never external writes.

Four pillars

Install the packs your systems need.

01

Govern

Run the AI management system: policies, accountability, inventory, risk and impact, data and suppliers, and the metrics that show systems are under control.

Workflows
33
Evaluations
7
Packs
5
  • AI governance foundationStart here: scope, policies, accountability and literacy.
  • Inventory and lifecycleRegister systems, control releases and changes, track cost and sustainability.
  • Risk and impactRisk criteria, assessment, treatment and impact assessment.
  • Data and privacyData provenance and quality, retention, privacy and residency.
  • Third parties, IP and liabilityVendors, licensing and liability.

02

Evaluate

Test systems before and after release: quality, fairness and bias (FairMind's core engines), agent behaviour and production drift.

Workflows
12
Evaluations
46
Packs
5
  • Quality evaluationPre-release quality, independent testing and data quality.
  • Fairness and biasFairMind's fairness engines.
  • Explainability and oversightExplanations and human oversight that works.
  • Production monitoringDrift, degradation and runaway detection in operation.
  • Use-case evaluation suitesHiring, credit, finance, health and more.

03

Secure

Protect LLM applications and agents against the OWASP, MITRE ATLAS and NIST adversarial threat classes, secure the AI supply chain, defend against offensive AI and deepfakes, and respond to incidents.

Workflows
29
Evaluations
53
Packs
8
  • Secure LLM applicationsOWASP LLM Top 10 for chat, RAG and copilots.
  • Secure agentic AIOWASP Agentic and MCP Top 10 for agents.
  • AI supply chainAI-BOM, model intake and scanning.
  • Adversarial ML and red teamingMITRE ATLAS and NIST adversarial ML coverage.
  • Offensive AI and deepfake defenceImpersonation, deepfakes and AI-driven fraud.
  • Content safetyHarmful content and misuse.
  • Incident responseHandle, contain, report and learn.
  • FairMind agent self-assuranceEvidence that FairMind's own agent is governed.

04

Comply

Meet specific legal and certification obligations (EU AI Act, ISO/IEC 42001, AIUC-1 and others) and show it, through journeys that add packs up to each framework.

Workflows
16
Evaluations
3
Packs
3
  • Regulatory obligationsApplicability, conformity, registration, transparency and reporting.
  • Audit and certification readinessWhat certification auditors sample.
  • Rights of affected peopleExplanation, appeal, consultation and accessibility.

109 evaluations

Tests for the systems you actually run. Thresholds declared first.

Every evaluation in the catalogue, with its metric, its threshold and where that threshold comes from. The fill shows what FairMind can run today, so you can see the gaps as well as the coverage.

  • Engine in FairMind 10
  • Partial engine 24
  • Specified, not built yet 75

Security 53

Behaviour and fairness 46

Governance and compliance 10

The agent is governed too

Twenty controls on our own agent.

Per-step scoped identity. A deterministic tool allowlist prompts cannot widen. Collected content treated as data, never instructions. A hash-chained trace of every call. A kill switch per run. You can see our agent's metrics and incidents; we cannot hide them.

Ledger deployment_release_gate

    Decide at the gate above. Your decision is hashed and chained here, on top of the run before it.

    1. collect14 items · scopes: read, short-lived
    2. check12 pass · 1 fail · 1 cannot conclude
    3. suggestgo 5% → 25% → all · confidence 0.78 to 0.84
    4. dry runblast radius 1 release · rollback path confirmed
    5. askapprover: AI system owner + Risk · releaser is not the approver
    6. actsigned approval posted to the CI/CD gate
    7. verifytarget re-read · match
    8. fileevidence → ISO/IEC 42001 8.1, AIUC-1 E004, NIST AI RMF MANAGE 1.1, EU AI Act Art 43
    9. tracesha256 e1b0…9a44 · chained to the previous run

    Questions teams ask

    What people ask before the pilot.

    What is agentic AI governance?

    Agentic AI governance uses AI agents to run the repeatable parts of governing AI systems: collecting evidence, checking it against framework requirements, proposing records and fixes, and filing the trace. The decisions stay with people. In FairMind, every workflow stops at an approval gate before anything is recorded or changed.

    What does a FairMind agent actually do?

    It runs a workflow made of seven step types: collect, check, suggest, ask a person, act, verify, and file evidence. Today the trigger, escalate and file steps run themselves, the gate steps open an approval request, and a person completes the collect, check, suggest, act and verify steps, which the ledger records as manual. Agent-executed steps arrive workflow by workflow, behind the same gates, with scoped read-only credentials, rules before model judgement, and a citation for every finding.

    What does FairMind evaluate?

    The catalogue defines 109 evaluations: 53 for security, 46 for behaviour and fairness, and 10 for governance and compliance. They cover agents (tool least privilege, out-of-scope actions, MCP permissions, kill switches), LLM apps and RAG (prompt injection, jailbreaks, prompt leakage, groundedness, harmful content), models and data (accuracy, calibration, drift, membership inference, backdoors) and decision systems (selection-rate parity, error-rate parity, intersectional slices). Each one names its metric, its threshold and where that threshold comes from, and every evaluator declares whether it is validated, runs through an external provider, reads metadata only, or is not yet available.

    Can the agent change things on its own?

    Only inside FairMind, and only if you set that workflow to autonomy level A3. The default is A2: the agent proposes and a person approves. External systems are never written without an approval. If an approver is unreachable or no policy covers the action, nothing happens.

    Which frameworks does FairMind cover?

    The registry holds 51 frameworks, including the EU AI Act, ISO/IEC 42001, NIST AI RMF and AI 600-1, the OWASP Top 10 for LLM and Agentic applications, AIUC-1, MITRE ATLAS, the OWASP MCP Top 10, CSA MAESTRO, Singapore’s Model AI Governance Framework, and national rules from India, Japan, Korea, China and US states. One workflow run counts toward every requirement it maps to.

    How is FairMind’s own agent governed?

    Twenty runtime controls apply to it: per-step scoped identity, a deterministic tool allowlist that prompts cannot widen, collected content treated as data rather than instructions, mandatory citations, pinned model and prompt versions with golden-set evaluations, a hash-chained immutable trace, cost and rate caps, and a kill switch per run. Customers can see the agent’s metrics and incidents.

    Is FairMind a certification?

    No. FairMind produces and organises evidence and records human decisions. Certification and conformity decisions stay with your reviewers, auditors, and, where relevant, accredited bodies.

    Let the agents do the paperwork. Keep the decisions.

    Request access Book a walkthrough