EU AI Act
A reference for relevant AI uses, including documentation, risk management, and human oversight where the Act’s high-risk provisions apply.
Applicability depends on your role, AI use, EU connection, and applicable provisions and dates.
Discuss an assessment AI GOVERNANCE, MADE OPERATIONAL
Turn scattered AI use into a governed system of record.
AI governance for insurers, lenders, and health organizations. Connect systems, owners, evidence, and action.
THE GOVERNANCE WORKSPACEINTERACTIVE PREVIEW
THREE SYSTEMS. ONE CONNECTED RECORD.
Open workspace See where AI is used, who owns the decision, and what supports it.
Arrange and filter the graph Scope · Grouping · Search
This arrangement is too dense to keep labels separate at the current zoom. Showing the complete record list. Zoom in or narrow the scope to return to the graph.
Every matching record is available in this list, including records collapsed in the graph.
Accountability, controls, and evidence. In one view.
FROM AI INVENTORY TO ONGOING OVERSIGHTANVAYLO
THE CONNECTED CAPABILITIES
Six capabilities.
One line of sight.
From an AI use case to the decision it supports. Follow the people, obligations and evidence in between.
Explore the capability mapAI & Agent Inventory
AI inventory demo · Agent records plannedStart with purpose, ownership and dependencies—not just a list of model names.
Policy & Obligation Mapping
Demonstration + planned scopeConnect requirements to policies and controls, with a recorded rationale for each relationship.
Decision & Evidence Graph
Demonstration + planned scopeBring policies, controls, contracts, evidence and transactions into one navigable record.
Risk & Review Workflows
Demonstration + planned scopeConnect risk assessment to accountable review, conditions, exceptions and the next action.
Vendor & Contract Assurance
Demonstration + planned scopeRelate supplier commitments and open questions to the AI systems and decisions that depend on them.
Oversight & Reporting
Demonstration + planned scopeConnect leadership reporting to the owners, evidence gaps and follow-up work underneath it.
The public workspace is an interactive demonstration. Production workflows are planned and scoped separately; no customer documents are ingested or monitored here.
See the phased roadmapTHE SYSTEM PROFILE
Inside an AI system.
Choose an AI system. Open any facet to see the people, decisions and records behind it.
A human underwriter reviews the submission summary and retains the underwriting decision.
THE GUIDED ASSESSMENT
Three stages of assessment. Four tangible outputs.
The planned assessment combines a scoped four-to-six-week engagement with twelve months of workspace access. Scope, availability and timing are confirmed in your proposal.
- 01
Frame the work
Agree the scope, identify stakeholders, and gather the starting inventory.
- 02
Build the record
Document systems and ownership, assess risk, and connect controls to evidence.
- 03
Set the next actions
Review observations, agree owners and next actions, and define the workspace handover.
THE OUTPUTS, NOT MORE STEPS
What you receive.
EXAMPLE OUTPUT
AI registry and risk observations
Systems, purposes, owners, model context and recorded risk considerations.
Open the working previewEXAMPLE OUTPUTEvidence and control gap pack
Bounded observations, supporting records and proposed review decisions.
Open the working previewEXAMPLE OUTPUTExecutive AI risk brief
A leadership summary with a scoped registry and source appendix.
Open the working previewEXAMPLE OUTPUTOngoing workspace and action register
Assigned work, status, priority, due dates and the evidence needed to close it.
Open the working previewCONTINUITY, BUILT IN
The intended outcome is an ongoing workspace with named owners and a maintained action register; secure customer access is planned.
- Planned engagement
- 4–6 weeks
- A defined assessment scope
- Fixed fee
- Planned workspace access
- 12 months
- Source handling confirmed before pilot
- Agreed scope
INDUSTRY EXAMPLES
Industry examples. Different work. Shared foundations.
Start with the same connected record. Shape the assessment around the decisions, workflows, and people in your organization.
From underwriting support to claims review.
Document how AI supports insurance workflows, who reviews its output, and which evidence supports the controls. Connect internal systems and vendor relationships in the same inventory.
ILLUSTRATIVE SCENARIO
An insurer introduces a submission assistant while another team tests a claims model.
THE WORKING OUTPUT
One inventory linking use cases to owners, review evidence, and outstanding vendor questions. Consider the NAIC AI Model Bulletin alongside relevant state insurance guidance.
Hypothetical workflows, not client case studies or measured outcomes.
A STARTING POINT FOR YOUR TEAM
Choose the decision
you need to support.
A supplier review. An AI rollout. A leadership question. Choose the decision you need to make.
Explore all six solutionsEstablish an AI Program
A new AI program lead or a lean risk team
Review AI Vendors
Procurement and third-party risk teams
Oversee High-Impact Decisions
Insurance, lending and health operations leaders with risk reviewers
Prepare an Assurance Review
Compliance, internal assurance and customer-facing trust teams
Plan an Agent Rollout
AI product owners, security teams and business sponsors
Brief Leadership
Executives, boards and governance committees
FRAMEWORK REFERENCES
Ground the conversation in recognized references.
Select relevant references when agreeing the assessment scope. These are reference points, not a claim of automated coverage or a determination of your obligations.
NIST AI RMF 1.0
A voluntary framework for structuring AI risk discussions and governance practices across the lifecycle, adaptable to different organizations, sectors, and AI use cases.
Voluntary guidance, not a legal compliance determination or endorsement.
ISO/IEC 42001
An international management system standard for organizing AI responsibilities, policies, risk processes, and continual improvement across organizations that develop, provide, or use AI.
Reference does not establish conformity or certification. Standard text is not reproduced here.
NAIC AI Model Bulletin
A model bulletin describing insurer AI governance expectations, including a written AI systems program, third-party oversight, consumer risk considerations, and documentation for regulatory review.
The model is not itself law or regulation. State adoption, wording, and applicability vary.
CLEAR WORKING BOUNDARIES
Your evidence. Human judgment.
A useful governance record makes its boundaries as clear as its findings.
Source handling is agreed first
The public demonstration uses bundled example records. Customer source handling, storage and any AI processing will be agreed before a secure pilot begins.
People own the decisions
The assessment documents observations and proposed actions. Your team reviews the findings and retains responsibility for decisions and risk acceptance.
Specialist questions go to specialists
Legal, accounting, actuarial, and certification questions belong with appropriately qualified professionals. Sector-specific applicability should be confirmed with qualified counsel.
START A CONVERSATION
What needs to be clearer about your AI?
Tell me what your organization is using, where ownership or evidence is unclear, and what you want the assessment to deliver.
Or email Anuj directlyanuj@anvaylo.com